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] [day] [month] [year] [list]
Message-ID: <20250902081014.0e6eaaf2.xunil@tahomasoft.com>
Date: Tue, 2 Sep 2025 08:10:14 -0400
From: Erik Beck <xunil@...omasoft.com>
To: Chukun Pan <amadeus@....edu.cn>
Cc: andrew@...n.ch, conor+dt@...nel.org, devicetree@...r.kernel.org,
 heiko@...ech.de, krzk+dt@...nel.org, linux-arm-kernel@...ts.infradead.org,
 linux-kernel@...r.kernel.org, linux-rockchip@...ts.infradead.org
Subject: Re: [PATCH 3/4] arm64: dts: rockchip: Add HINLINK H68K

On Tue,  2 Sep 2025 15:00:19 +0800
Chukun Pan <amadeus@....edu.cn> wrote:

> Hi,

Greetings!

> 
> > So what I said earlier regarding rgmii vs rgmii-id DOES NOT hold up
> > under more rigorous and careful testing.  
> 
> So the following question does not exist?

Yes, that is correct. My earlier question/comment/concern about a huge speed
difference between rgmii-id mode and rmgmii doesn't seem to exist when more
careful experiments and testing methods are used. I'll elaborate some more
below. And I apologize for any inconvenience. 

> >> Changing this makes a huge difference in the ethernet throughput speed.
> >> With rgmii-id mode specified, throughput is about 6.5 Mbs. Changing this
> >> to rgmii mode increases throughput to about 960 Mbs.  
> 

I'm still a bit surprised that my first test method (web-based, like fast.com
or self-hosted internal download speeds) gave such different results from more
comprehensive and careful testing with iperf3. While certainly web-based
methods can't target a particular interface, nor do they have many other
potential adjustments or refinements (UDP vs TCP, etc), they have been and
continue to be a useful tool for me to get quick information about the status
and health of my network and its devices. 


> If the iperf3 test does not reach Gigabit, you can run it in
> multiple threads. e.g. `iperf3 -c xxx -P 4`

Thanks for the tip.

I remain keenly interested in seeing a patch for the Hinlink/LinkStar devices
make it into the mainline kernel. Please let me know how I can help your
(and others) efforts to accomplish that.

Regards,

Erik


Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ