Request: tester for ipq8064 soc

For Developers
82

i've tested with merging the whole branch on an EA8500 and it boots ok
there's a log in the thread of the whole boot up including u-boot from serial

I got stuck on trying to get multiple vlans working on a port but I admit I haven't tried just making vlan interfaces off the lanX

If you want me to test on ipq40xx as well I can. I tried out LGA1150/exp branch and it's ok except the patch that enables pstore prevents the kernel from booting so I had to revert that. There's also a patch there that removes the eth1 interface but it means there's a period of time on bootup where all the ports are bridged so lan devices can end up grabbing the internet ip.

83

gives me the idea of possibly testing with this reverted:

84

consider that since dsa doesn't support multiple cpu port, in the current implementation the first cpu port is always used.

If we find a good implementation of multi cpu port for dsa, i can make the required changes to add support for this with dsa... (additional bit are required to set the secondary cpu port to accept packet)

85

oh I see, no point reverting that since your tree already comments out the mdio0 section

should I comment out this section in the EA8500 dts ?

&mdio0 {
        phy4: ethernet-phy@4 {
                reg = <4>;
        };
};
86

yes didn't notice that

87

I am going to build and flash it now :wink:

I didn't notice any sudden disconnection these days...
This is the dmesg section of yesterday:

[337384.854216] qca8k 37000000.mdio-mii:10 lan2: Link is Up - 1Gbps/Full - flow control rx/tx
[337384.854324] br-lan: port 2(lan2) entered blocking state
[337384.861545] br-lan: port 2(lan2) entered forwarding state
[365290.188721] qca8k 37000000.mdio-mii:10 lan2: Link is Down

365290−337384 = 27906
27906/60/60 = ~8 hours of connection (it's right)

Is there any chance that a fast disconnection can't be detected by the driver?

88

Problem is that sometimes the port goes down and up without a reason... found the cause in the mdio driver that it should be good now... but still some lan port still have this problem (don't know if it's caused but the other device connected)
anyway i have almost 3 days of wan connection up so could be some eee protocol that cause some problem with the switch (in the original driver any eee is disabled since it does cause stability problem)

89

My PPPoE is still up after 5 days :wink:
image

90

yes that could be a hint that the lan port drop because some energy efficient protocol are applied to the port.

93

Think we can move here... I moved all the testing commit to the kernel 5.10 bump...

94

Good job! I'm not at home this week but I'll try to upgrade my C2600 remotely to test your PR. Settings should be the same as your last DSA version right?

95

mhhh honestly i would suggest you to remote upgrade it only if you really can... (so only if the device is not critical and IT'S NOT A PROBLEM IF IT HAS WHATEVER PROBLEM THAT WOULD CAUSE ANY MALFUNCTION)

96

My C2600 is a dumb router, so not critical. Anyway, it's now upgraded and fully functional.

97

good to know... Can you confirm it's kernel 5.10 ? Also can you post a logread?

98

Yes, it's a C2600 on kernel 5.10. I've just pm'd you the log. Cheers.

99

@rog thx for the log... it looks clean.

Anyway i think i will close this and move to the 5.10 testing since they have the same patches + new kernel.

Again thx a lot to everyone.

2 Likes
100

I flashed with your 5.10 for my c2600 and there is no wireless tab...
I reflashed, rebooting.. still no wireless option in network

101

can you give me a logread?

102

the regular syslog?

log

what I just did,

rm -f /etc/config/wireless

wifi config > /etc/config/wireless

with that, wireless appeared in winscp but it is completely empty and still not there in LuCi

103

I've seen the same thing with @ACwifidude build some months ago. I sysupgraded with the same build and it was ok.