MIME-Version: 1.0
Message-ID: <trinity-dd260791-3637-4193-8f93-a9fcdb013dcb-1676722705920@3c-app-gmx-bap53>
From:   Frank Wunderlich <frank-w@public-files.de>
To:     =?UTF-8?Q?Ar=C4=B1n=C3=A7_=C3=9CNAL?= <arinc.unal@arinc9.com>
Cc:     Rob Herring <robh+dt@kernel.org>,
        Krzysztof Kozlowski <krzysztof.kozlowski+dt@linaro.org>,
        Matthias Brugger <matthias.bgg@gmail.com>,
        AngeloGioacchino Del Regno 
        <angelogioacchino.delregno@collabora.com>,
        devicetree@vger.kernel.org, linux-kernel@vger.kernel.org,
        linux-arm-kernel@lists.infradead.org,
        linux-mediatek@lists.infradead.org, erkin.bozoglu@xeront.com
Subject: Aw: Re:  Re: [PATCH v3 0/5] arm: dts: mt7623: relocate gmacs,
 mt7530 switch, and add port@5
Content-Type: text/plain; charset=UTF-8
Date:   Sat, 18 Feb 2023 13:18:25 +0100
Importance: normal
Sensitivity: Normal
In-Reply-To: <27a26da8-8297-5327-7493-54d8359b6970@arinc9.com>
References: <20230210182505.24597-1-arinc.unal@arinc9.com>
 <c3ab9a9b-3eb2-8fb0-d5d7-c0b7c684d3a7@arinc9.com>
 <trinity-dab715b9-3953-40da-bc25-c4c2a5e9b7c3-1676715866453@3c-app-gmx-bap53>
 <27a26da8-8297-5327-7493-54d8359b6970@arinc9.com>
Content-Transfer-Encoding: quoted-printable
UI-OutboundReport: notjunk:1;M01:P0:VZV6dh1A5Ag=;Buw383pPwakwuXb/1j7tqGijIUR
 3iem6oEThCQ7VPv417sYU4cqM8eTKTxf9CiPBWZj0UVIEx5YqUC6qyjWfwzTKiFG9DV6fsCyW
 1L08OdWZk/WFkwhx+Yzszcl4xykEjrxGIe9VEaLQjGrSvQwvipUtUGKrLdz5zvz7RLpui7KEZ
 hZnyJoOlgB2HZPMz64EMhCfSHtraXEtk2yWOYPlaGFs+eN4BpBfsv6zpW+xJoN/dG3dAiagpK
 XeNygMNdfrW/l6umEVQpTfo2TKRkWsnnHbX9EVLLS4/7WfMsrh29cuG992KgIv0FS+BuNmnih
 cP3U/caOvlESKH1WGzpRzgavmcshBpvwPpNnjAepRmcg0iV28gm2yNefK/e5qVb4JbywWffsQ
 HD7xWCYpEl3gadxP8CeL3z0mdS0HlTFtjFEel0m4M2czy0XfrMNRFSSQ2gGafpbb11hHFaN1j
 g+eYQ1nep7Azz9tLxOzbjqPVZYCoEL4zrNU8lQ2MMl2gNJt9iOam7b55cwWcw4dob6OywGJwn
 6NS7i67/Xxmo7Jc46EfAUfXknUUZ8Pi27MFTruBIJnJNXr6kuG8Filbkqx8GZ3Sx5h6gWm8m5
 Pv2+qw9lTrqAksnDCEdeiYPqf2xekLAy7qpASheFkUDT65rPp+0wgeMqzkyNnaLXVPK6TY0AJ
 1eXEIIuRuW8O+KX3RbZm1XYUFWXD6w54jjw3hE4aRUr8pphpjd4PUKe3ofQ4ZwFe+yiyblyZa
 aLIMOsTfgmS7DeEwmowaJLIe+FPwC17LO5SS3OHxkYjh3G9mHGW8jhApjrtQU+/FCI0r1Jqij
 yf1y7MbtJcYwQvuRzjjL4J1wNpgKglbhp6PlrKMdzew3M=
Precedence: bulk

Hi,

> Gesendet: Samstag, 18=2E Februar 2023 um 11:49 Uhr
> Von: "Ar=C4=B1n=C3=A7 =C3=9CNAL" <arinc=2Eunal@arinc9=2Ecom>
> An: "Frank Wunderlich" <frank-w@public-files=2Ede>

> On 18=2E02=2E2023 13:24, Frank Wunderlich wrote:
> > Hi,
> >=20
> > finally get some time to bootup with this series on my r2=2E
> >=20
> > I see that inserting the port@5 as cpu-port maps this as default-cpu b=
ecause dsa-core uses first found cpu-port as
> > default (dsa_tree_setup_cpu_ports in dsa=2Ec) and because of lower ban=
dwidth (rgmii instead of trgmii) not the best choice=2E
> >=20
> > But it look worse=2E=2E=2Enetwork is currently broken (set both gmacs =
up)=2E
> > I see arp-packets reaching remote side, but reponse is not received by=
 r2
> >=20
> > I have tested it on 6=2E2-rc8 (wan-port), maybe additional patches are=
 needed?=2E=2E=2Euserspace setup should be right=2E
> >=20
> > so i added series on top of net-next (no additional patches except som=
e basic like build-script,defconfig and such)=2E=2E=2Esame result=2E=2E=2E
> >=20
> > i'm not sure if i change the mapping from userspace back to eth0, so d=
isabled port@5 in switch, now port6 is
> > cpu-port again and it works=2E=2E=2Eso something is wrong with port5 o=
f switch or gmac1=2E
>=20
> That's a driver issue and will be fixed once an accepted version of=20
> these patches [0] [1] [2] are applied to net-next=2E You should have the=
m=20
> on your inbox, I specifically told Richard to CC you=2E

yes, i've got them, but not applied when starting these tests=2E Not thoug=
ht, that this change is necessary=20
as we use both gmac on mt7531/bpi-r3 with out problems=2E

with these 3 Patches network-connection works, but only at ~624 Mbit/s in =
TX-Mode (started iperf3-client on r2 without -R) and massive retransmitts o=
n first run, other direction is clean=2E

and these Patches need to be applied first (when fixed up) else network is=
 broken=2E

> This is devicetree bindings=2E We're here to describe the hardware=2E Th=
e=20
> way a driver works should not affect describing the hardware=2E

thats right, but it should not break/change behaviour like now all ports h=
ave only ~600Mbit because of
moving them all to the other gmac which has obvious issues=2E

> To address the lower bandwidth situation you mentioned, a devicetree=20
> property to designate a CPU port as the default CPU port could be=20
> introduced=2E I'm not aware of a similar conversation so I'll send a mai=
l=20
> to netdev to discuss this=2E Will CC you=2E

isn't there a way to leave ports by default on the the better gmac (gmac0=
=3Dtrgmii)?
maybe moving port5 below port6=2E=2E=2Enot nice, but then port6 is the fir=
st cpu found=2E
or maybe defined the default cpu in driver which gets picked up by core (d=
efine port6 if available as default)=2E
Imho additional dts-propperty is wrong approch=2E=2E=2Eit should be handle=
d by driver=2E But cpu-port-selection is=20
currently done in dsa-core which makes it a bit difficult=2E

I'm not sure how the multi-cpu support in dsa-core ended and how you use t=
he other gmac not used by dsa-core

set master in userspace-config? i remember you've sent a patch adding call=
back for it=2E

imho dts change should be applied if these points are cleared=2E=2E=2E

regards Frank