> If I didn't know that though, my expectations would be that the
> Fair Queue packet scheduler would behave just like its name says, fair. :P
>
Only if the flows have same characteristics. This cannot be ensured for
hours long flows. This should be obvious.
If one flow is not trying to push en
Greetings fellow hackers :)
I have managed to reproduce my streaming unfairness with the intel e1000 driver.
I believe that commit
605ad7f184b6 ("tcp: refine TSO autosizing")
changed the behavior for 100 Mbps NIC:s with TSO enabled.
I have 2 e1000 cards connected in the same computer, with a net
On Wed, 2015-11-25 at 13:47 +0100, Niklas Cassel wrote:
> (SO_MAX_PACING_RATE would require modifying a lot of different user-space
> programs, so that is not really an option :( )
You can set a max rate on the packet scheduler itself, for all flows.
tc qdisc replace dev eth0 root fq maxrate 30M
On 11/20/2015 07:16 PM, Eric Dumazet wrote:
> On Fri, 2015-11-20 at 16:33 +0100, Niklas Cassel wrote:
>
>> I've been able to reproduce this on a ARMv7, single core, 100 Mbps NIC.
>> Kernel vanilla 4.3, driver has BQL implemented, but is unfortunately not
>> upstreamed.
>>
>> ethtool -k eth0
>> Of
On Fri, 2015-11-20 at 16:33 +0100, Niklas Cassel wrote:
> I've been able to reproduce this on a ARMv7, single core, 100 Mbps NIC.
> Kernel vanilla 4.3, driver has BQL implemented, but is unfortunately not
> upstreamed.
>
> ethtool -k eth0
> Offload parameters for eth0:
> rx-checksumming: off
> t
On 11/09/2015 05:07 PM, Eric Dumazet wrote:
> On Mon, 2015-11-09 at 16:53 +0100, Niklas Cassel wrote:
>> On 11/09/2015 04:50 PM, Eric Dumazet wrote:
>>> On Mon, 2015-11-09 at 16:41 +0100, Niklas Cassel wrote:
I have a ethernet driver for a 100 Mbps NIC.
The NIC has dedicated hardware for
On 11/09/2015 06:23 PM, Eric Dumazet wrote:
> On Mon, 2015-11-09 at 17:50 +0100, Niklas Cassel wrote:
>
>>
>> for i in `seq 1 20`; do ss -temoi dst 192.168.0.141; sleep 1; done
> ...
>> ESTAB 0 0192.168.0.1:54578
>> 192.168.0.141
On Mon, 2015-11-09 at 17:50 +0100, Niklas Cassel wrote:
>
> for i in `seq 1 20`; do ss -temoi dst 192.168.0.141; sleep 1; done
...
> ESTAB 0 0192.168.0.1:54578
> 192.168.0.141:5201 rto:0.2 ato:0.04 cwnd:10 uid:20283
> ino:7
On 11/09/2015 05:24 PM, Eric Dumazet wrote:
> On Mon, 2015-11-09 at 08:07 -0800, Eric Dumazet wrote:
>
>> Your numbers suggest a cwnd growth then, which might show a CC bug.
>>
>> Please run the following when your iper3 runs on regular 4.3 kernel
>>
>> for i in `seq 1 10`
>> do
>> ss -temoi dst 1
On Mon, 2015-11-09 at 08:07 -0800, Eric Dumazet wrote:
> Your numbers suggest a cwnd growth then, which might show a CC bug.
>
> Please run the following when your iper3 runs on regular 4.3 kernel
>
> for i in `seq 1 10`
> do
> ss -temoi dst 192.168.0.141
> sleep 1
> done
>
Another thing to tr
On Mon, 2015-11-09 at 16:53 +0100, Niklas Cassel wrote:
> On 11/09/2015 04:50 PM, Eric Dumazet wrote:
> > On Mon, 2015-11-09 at 16:41 +0100, Niklas Cassel wrote:
> >> I have a ethernet driver for a 100 Mbps NIC.
> >> The NIC has dedicated hardware for offloading.
> >> The driver has implemented TSO
On 11/09/2015 04:50 PM, Eric Dumazet wrote:
> On Mon, 2015-11-09 at 16:41 +0100, Niklas Cassel wrote:
>> I have a ethernet driver for a 100 Mbps NIC.
>> The NIC has dedicated hardware for offloading.
>> The driver has implemented TSO, GSO and BQL.
>> Since the CPU on the SoC is rather weak, I'd rat
On Mon, 2015-11-09 at 16:41 +0100, Niklas Cassel wrote:
> I have a ethernet driver for a 100 Mbps NIC.
> The NIC has dedicated hardware for offloading.
> The driver has implemented TSO, GSO and BQL.
> Since the CPU on the SoC is rather weak, I'd rather
> not increase the CPU load by turning off off
13 matches
Mail list logo
