[pve-devel] playing with 3.10

Stefan Priebe - Profihost AG s.priebe at profihost.ag
Fri Aug 30 09:34:22 CEST 2013


Am 30.08.2013 09:31, schrieb Alexandre DERUMIER:
>>> Yes but i can't use that as gvrp is not working in this case. The kernel
>>> does not announce gvrp continuously if you use another bridge on top of
>>> the bond or eth.
> 
> So, maybe is it a regression with vmbr1.3021 at vmbr1 setup?  (It's not the first time that things break in bridge code, it's a little mess)
> 
> Don't known if you can test with tag ethX to see if it's working?

This works fine on the host:
ip link add link vmbr1 name vmbr1.3025 type vlan id 3025 gvrp on
reorder_hdr on

ip a add 192.168.99.10/24 bro 192.168.99.255 dev vmbr1.3025
ip link set vmbr1.3025  up

But i'm not able to get a tap device working.

Stefan

> ----- Mail original ----- 
> 
> De: "Stefan Priebe - Profihost AG" <s.priebe at profihost.ag> 
> À: "Alexandre DERUMIER" <aderumier at odiso.com> 
> Cc: pve-devel at pve.proxmox.com 
> Envoyé: Vendredi 30 Août 2013 09:22:20 
> Objet: Re: [pve-devel] playing with 3.10 
> 
> Am 30.08.2013 05:43, schrieb Alexandre DERUMIER: 
>>>> The relevant commit says it's disabled by default. Strangely i don't see 
>>>> how to enable and use this from the commits message. 
>> yed, I see that too. 
>>
>>
>> also found this: 
>>
>> "Bridge VLAN kernel/iproute2 incompatibility" 
>> http://www.spinics.net/lists/linux-ethernet-bridging/msg04942.html 
>>
>> don't known if if could be related? a kernel patch has been submit in august. 
> 
> I saw that too but this should be only relevant when using the new HW 
> filter. 
> 
>>>> mhm i could try that. So you mean instead of adding: 
>>>> vmbr1.3021 at vmbr1 
>>>> and 
>>>> vmbr1v3021 
>>>>
>>>> just use vmbr1.3021 at vmbr1 and add the tap device to that one? 
>>
>> mmm, what is vmbr1.3021 at vmbr1? 
>> current proxmox implementation is ethx.302---vmbr1v302 
>> (can you do a brctl show?) 
> Yes but i can't use that as gvrp is not working in this case. The kernel 
> does not announce gvrp continuously if you use another bridge on top of 
> the bond or eth. 
> 
>> But, with new vlan code, what we will do should be 
>> ethx----vmbr1----tap 
>> and configure virtual port on vmbr1, where tap is plugged, with vlan tag 302. (like a real switch) 
> 
> Right now i'm trying to use the old variant as the new iproute2 package 
> is also only available in sid. 
> 
> Stefan 
> 
>>
>> ----- Mail original ----- 
>>
>> De: "Stefan Priebe" <s.priebe at profihost.ag> 
>> À: "Alexandre DERUMIER" <aderumier at odiso.com> 
>> Cc: pve-devel at pve.proxmox.com 
>> Envoyé: Jeudi 29 Août 2013 20:58:04 
>> Objet: Re: [pve-devel] playing with 3.10 
>>
>> Am 29.08.2013 20:52, schrieb Alexandre DERUMIER: 
>>> Hi Stefan, 
>>>
>>> I don't have a look a vlan-bridge implementation in new kernel. 
>>>
>>> As far I known, bridge support now vlan tagging (by port?). Maybe it's tagging by default ? 
>>
>> The relevant commit says it's disabled by default. Strangely i don't see 
>> how to enable and use this from the commits message. 
>>
>>> Mayve try to configure vlan1 to disable vlan filtering ? 
>> What do you mean by that? 
>>
>>> Or maybe can we try to implement new vlan code ? 
>>> (don't tag anymore on eth with multibridge, but use only 1 bridge with tagging port) 
>> mhm i could try that. So you mean instead of adding: 
>> vmbr1.3021 at vmbr1 
>> and 
>> vmbr1v3021 
>>
>> just use vmbr1.3021 at vmbr1 and add the tap device to that one? 
>>
>>> I really don't known how it's work, I'll try to have a look at it next week. 
>>> (I think rhel7 kernel will use it, so it should come next year with openvz for proxmox) 
>> Yes it will use that one. 
>>
>> Stefan 
>>
>>>
>>> ----- Mail original ----- 
>>>
>>> De: "Stefan Priebe - Profihost AG" <s.priebe at profihost.ag> 
>>> À: pve-devel at pve.proxmox.com 
>>> Envoyé: Jeudi 29 Août 2013 16:02:12 
>>> Objet: [pve-devel] playing with 3.10 
>>>
>>> Hi, 
>>>
>>> as Linux Kernel 3.10 ist the newest long term kernel i'll start 
>>> migrating all my machines to it. 
>>>
>>> Everything works fine except for tap devices with VLANs on top of bridges. 
>>>
>>> I'm pretty sure that this i related to this commit: 
>>> https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/commit/?id=93197b 
>>>
>>> Has any body an idea how to get it working again for proxmox? I already 
>>> tried to compile the kernel with or without BRIDGE_VLAN_FILTERING but it 
>>> doesn't make a difference. VLAN packages never passes the bond to bridge 
>>> to the tap device. 
>>>
>>> Greets, 
>>> Stefan 
>>> _______________________________________________ 
>>> pve-devel mailing list 
>>> pve-devel at pve.proxmox.com 
>>> http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel 
>>>



More information about the pve-devel mailing list