Thread (9 messages) flat view 9 messages, 4 authors, 2021-06-09

RE: [PATCH V1 net-next 0/2] net: fec: fix TX bandwidth fluctuations

From: Joakim Zhang <hidden>
Date: 2021-06-09 12:04:12
Also in: lkml

Hi Frieder,
-----Original Message-----
From: Frieder Schrempf <redacted>
Sent: 2021年6月8日 22:19
To: Joakim Zhang <redacted>; Andrew Lunn
[off-list ref]
Cc: davem@davemloft.net; kuba@kernel.org; netdev@vger.kernel.org;
linux-kernel@vger.kernel.org; dl-linux-imx [off-list ref]
Subject: Re: [PATCH V1 net-next 0/2] net: fec: fix TX bandwidth fluctuations

Hi Joakim, hi Andrew,

On 08.06.21 05:23, Joakim Zhang wrote:
quoted
Hi Frieder,
quoted
-----Original Message-----
From: Andrew Lunn <andrew@lunn.ch>
Sent: 2021年5月27日 22:06
To: Joakim Zhang <redacted>
Cc: davem@davemloft.net; kuba@kernel.org;
frieder.schrempf@kontron.de; netdev@vger.kernel.org;
linux-kernel@vger.kernel.org; dl-linux-imx [off-list ref]
Subject: Re: [PATCH V1 net-next 0/2] net: fec: fix TX bandwidth
fluctuations

On Thu, May 27, 2021 at 12:10:47PM +0000, Joakim Zhang wrote:
quoted
Hi Frieder,

As we talked before, could you please help test the patches when you
are
free? Thanks.

Hi Frieder

If you can, could you also test it with traffic with a mixture of VLAN
priorities.
quoted
quoted
You might want to force the link to 10Full, so you can overload it.
Then see what traffic actually makes it through.
Did your mailbox get bombed, let you miss this mail, hope you can see this
reply.
quoted
Could you please give some feedback if it is possible? Thanks :-)
Thanks for the patches! As far as the bandwidth drops are concerned, this
solves the problem. According to my simple iperf tests the untagged TX traffic
now always goes to queue 0 and therefore doesn't see any random bandwidth
limitations anymore.

Regarding Andrew's request for testing this with some tagged traffic mix, my
problem is that though I have used VLANs before and know how to set them up,
I've never done anything with QoS, so I don't really now how to set the
priorities (looks like I need to set up internal priorities and egress mapping
somehow!?). If you have any pointers for this it would be appreciated. I
probably could do some quick verification tests, but I don't have the time to
really dive into the topic.
Thanks a lot for your testing! Could I add your t-b and r-b tag when I repost the patches to change functions into static as Jakub commented before?

Best Regards,
Joakim Zhang
Best regards
Frieder
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help