[pve-devel] bridge vlan range, kernel 4.1 : "message truncated" warning when too much vlans defined

[Alexandre DERUMIER]

Thanks Ronen,

I have tested

- char buf[16384]; 
+ char buf[3*16384]; 


and It's working fine to handle 4096 vlans numbers.

(The problem was really the number of vlans ids defined, 
could be 1 interface with 2-4095vlans, or 100 interfaces with same 2-4095vlans).

With this buffer size it's working fine


Thanks for the fast reply.

Regards,

Alexadre

----- Mail original -----
De: "Arad, Ronen" <ronen.arad at intel.com>
À: "aderumier" <aderumier at odiso.com>, "netdev" <netdev at vger.kernel.org>
Cc: "pve-devel" <pve-devel at pve.proxmox.com>
Envoyé: Jeudi 30 Juillet 2015 04:03:06
Objet: RE: bridge vlan range, kernel 4.1 : "message truncated" warning when too much vlans defined

The easiest fix is in iproute2. 
Increasing the buffer size used in rtnl_dump_filter_l 3x empirically resolves the 
issue. 

The root-cause is due to the buffer size calculation of dump request in the 
kernel. It is based on the total number of VLANs and does not account for 
compressed vlan requests. 
The flag argument is not propagated to get size functions. 

Fixing the size calculation in the kernel would work most of the time but would 
Fail when the number of ranges (and/or non-contiguous VLANs) exceeds ~1800. 

Therefore, fixing the size calculation in the kernel does not worth the 
Effort. Fixing iproute2 seems the easy way to overcome the issue. 

diff --git a/lib/libnetlink.c b/lib/libnetlink.c 
index 901236e..25c1c02 100644 
--- a/lib/libnetlink.c 
+++ b/lib/libnetlink.c 
@@ -198,7 +198,7 @@ int rtnl_dump_filter_l(struct rtnl_handle *rth, 
.msg_iov = &iov, 
.msg_iovlen = 1, 
}; 
- char buf[16384]; 
+ char buf[3*16384]; 
int dump_intr = 0; 
iov.iov_base = buf; 

Ronen 

>-----Original Message----- 
>From: netdev-owner at vger.kernel.org [mailto:netdev-owner at vger.kernel.org] On 
>Behalf Of Alexandre DERUMIER 
>Sent: Wednesday, July 29, 2015 11:25 AM 
>To: netdev 
>Cc: pve-devel 
>Subject: bridge vlan range, kernel 4.1 : "message truncated" warning when too 
>much vlans defined 
> 
>Hi, 
> 
>I'm currently testing vlan range on bridge filtering with kernel 4.1. 
> 
> 
>If I defined too much vlans, or too big vlan range, 
>I have a warning "message truncated", even if I'm using "-c" 
> 
> 
># bridge -c vlan 
>Message truncated 
>port vlan ids 
>eth2 1 PVID Egress Untagged 
> 
>bond0 1 PVID Egress Untagged 
>2-1850 
> 
>vmbr0 1 PVID Egress Untagged 
>94 
> 
> 
>or 
># bridge -c vlan 
>Message truncated 
>port vlan ids 
>eth2 1 PVID Egress Untagged 
>2-900 
>bond0 1 PVID Egress Untagged 
>2-900 
> 
>vmbr0 1 PVID Egress Untagged 
>94 
> 
> 
> 
>Also, more important, this is also impact "ip", and truncate the output 
> 
>#ip link 
> 
>Message truncated 
>1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN mode 
>DEFAULT group default 
>link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 
>2: eth2: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 9000 qdisc pfifo_fast master 
>vmbr1 state DOWN mode DEFAULT group default qlen 1000 
>link/ether 00:15:17:80:29:e6 brd ff:ff:ff:ff:ff:ff 
>3: eth3: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN mode DEFAULT 
>group default qlen 1000 
>link/ether 00:15:17:80:29:e7 brd ff:ff:ff:ff:ff:ff 
>4: eth0: <BROADCAST,MULTICAST,SLAVE,UP,LOWER_UP> mtu 1500 qdisc mq master 
>bond0 state UP mode DEFAULT group default qlen 1000 
>link/ether 00:1a:a0:3c:98:c5 brd ff:ff:ff:ff:ff:ff 
>5: eth5: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc pfifo_fast master 
>ovs-system state UP mode DEFAULT group default qlen 1000 
>link/ether 68:05:ca:09:c3:7c brd ff:ff:ff:ff:ff:ff 
>6: eth4: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc pfifo_fast master 
>ovs-system state UP mode DEFAULT group default qlen 1000 
>link/ether 68:05:ca:09:c3:7d brd ff:ff:ff:ff:ff:ff 
>7: eth1: <BROADCAST,MULTICAST,SLAVE,UP,LOWER_UP> mtu 1500 qdisc mq master 
>bond0 state UP mode DEFAULT group default qlen 1000 
>link/ether 00:1a:a0:3c:98:c5 brd ff:ff:ff:ff:ff:ff 
>8: ovs-system at NONE: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN mode 
>DEFAULT group default 
>link/ether 7e:dd:5d:ea:ef:26 brd ff:ff:ff:ff:ff:ff 
>9: vmbr10 at NONE: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc noqueue state 
>UNKNOWN mode DEFAULT group default 
>link/ether 68:05:ca:09:c3:7d brd ff:ff:ff:ff:ff:ff 
>10: bond1 at NONE: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc noqueue state 
>UNKNOWN mode DEFAULT group default 
>link/ether 8a:c3:da:b0:9d:87 brd ff:ff:ff:ff:ff:ff 
>11: bond0 at NONE: <BROADCAST,MULTICAST,MASTER,UP,LOWER_UP> mtu 1500 qdisc 
>noqueue master vmbr0 state UP mode DEFAULT group default 
>link/ether 00:1a:a0:3c:98:c5 brd ff:ff:ff:ff:ff:ff 
>12: vmbr0 at NONE: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state 
>UP mode DEFAULT group default 
>link/ether 00:1a:a0:3c:98:c5 brd ff:ff:ff:ff:ff:ff 
>13: vmbr0.94 at vmbr0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue 
>state UP mode DEFAULT group default 
>link/ether 00:1a:a0:3c:98:c5 brd ff:ff:ff:ff:ff:ff 
> 
>---->should have more interfaces here 
> 
> 
>Is it a known bug ? 
> 
> 
> 
>Regards, 
> 
>Alexandre Derumier 
> 
>-- 
>To unsubscribe from this list: send the line "unsubscribe netdev" in 
>the body of a message to majordomo at vger.kernel.org 
>More majordomo info at http://vger.kernel.org/majordomo-info.html