it compiles without errors, boots without crash/oops, but it still does not work…i only see eth0 and eth1, but none of the external (USER-) ports in “ip a”
also nothing obviously in dmesg
root@bpi-r2-ubuntu:~# dmesg | grep -i 'dsa\|net'
[ 0.195904] NET: Registered protocol family 16
[ 0.468627] NET: Registered protocol family 2
[ 0.513491] NET: Registered protocol family 1
[ 1.068059] drivers/net/ethernet/mediatek/mtk_eth_soc.c:mtk_mdio_init[413]0 237bf14f
[ 1.075784] mtk_soc_eth 1b100000.ethernet: generated random MAC address 5e:32:b8:eb:72:9b
[ 1.084466] mtk_soc_eth 1b100000.ethernet: connected mac 0 to PHY at fixed-0:00 [uid=00000000, driver=Generic PHY]
[ 1.095246] mtk_soc_eth 1b100000.ethernet eth0: mediatek frame engine at 0xe14c0000, irq 228
[ 1.103682] mtk_soc_eth 1b100000.ethernet: generated random MAC address 12:a6:6c:0e:bd:20
[ 1.112250] mtk_soc_eth 1b100000.ethernet: connected mac 1 to PHY at fixed-0:01 [uid=00000000, driver=Generic PHY]
[ 1.122983] mtk_soc_eth 1b100000.ethernet eth1: mediatek frame engine at 0xe14c0000, irq 228
[ 1.489242] NET: Registered protocol family 17
[ 3.678471] NET: Registered protocol family 10
i also wrote to john crispin he will look over it…but maybe anyone here see’s the problem.
main challenge was that dsa2.c changes too much for 4.15 (not only functionnames or parameters changed). here i’m unsure if i included new code in right context…
seems that i got it working with some help of a DSA-Maintainer
branch is currently based on rc1 but i will import changes to 4.19 if i removed unnecessary stuff (started with printks) and changed mt7530_probe back to mdio_device (platform_device should not be used because it will be dropped in future, but i need to know how to link to mdio-node)
all interfaces are visible, wan is bound to eth1 and lan-ports to eth0…quick-tested wan and lan0 via ping. i can also deactivate the corresponding cpu-port (eth1 for wan or eth0 for lan0) to disable ping…so routing seems correct.
wonder why this is not used
it gets overridden
dsa_tree_setup_switches => dsa_port_setup => dsa_user_parse
sets ds->ports[index].upstream = cpu_port_index;
but dp->cpu_dp is not changed again as far as i see, but in slave.c master is changed from cpu_dp.master to ports[x].ethernet so only this seems to be used and cpu_dp ignored
had also removed clock-patch and in_atomic (with kernel-hacking) in 4.20-gmac for testing (try to get it mainline)
it seems that it does not have any effect
root@bpi-r2:~# iperf -s
------------------------------------------------------------
Server listening on TCP port 5001
TCP window size: 128 KByte (default)
------------------------------------------------------------
[ 4] local 192.168.0.42 port 5001 connected with 192.168.0.21 port 39930
[ ID] Interval Transfer Bandwidth
[ 4] 0.0-10.0 sec 920 MBytes 771 Mbits/sec
^Croot@bpi-r2:~# iperf -c 192.168.0.21
------------------------------------------------------------
Client connecting to 192.168.0.21, TCP port 5001
TCP window size: 43.8 KByte (default)
------------------------------------------------------------
[ 3] local 192.168.0.42 port 59860 connected with 192.168.0.21 port 5001
[ ID] Interval Transfer Bandwidth
[ 3] 0.0-10.0 sec 1.09 GBytes 940 Mbits/sec
root@bpi-r2:~# ifdown wan
[ 1445.764386] mt7530 mdio-bus:00 wan: Link is Down
root@bpi-r2:~# [ 1456.498533] mt7530 mdio-bus:00 lan0: Link is Up - 1Gbps/Full f
[ 1456.505837] IPv6: ADDRCONF(NETDEV_CHANGE): lan0: link becomes ready
root@bpi-r2:~# ip addr add 192.168.0.15/24 dev lan0
root@bpi-r2:~# uname -a
Linux bpi-r2 4.20.0-rc1-bpi-r2-gmac #237 SMP Thu Nov 29 10:06:38 CET 2018 armv7x
root@bpi-r2:~# iperf -s
------------------------------------------------------------
Server listening on TCP port 5001
TCP window size: 128 KByte (default)
------------------------------------------------------------
[ 4] local 192.168.0.15 port 5001 connected with 192.168.0.21 port 54590
[ ID] Interval Transfer Bandwidth
[ 4] 0.0-10.0 sec 861 MBytes 721 Mbits/sec
root@bpi-r2:~# iperf -c 192.168.0.21
------------------------------------------------------------
Client connecting to 192.168.0.21, TCP port 5001
TCP window size: 43.8 KByte (default)
------------------------------------------------------------
[ 3] local 192.168.0.15 port 44502 connected with 192.168.0.21 port 5001
[ ID] Interval Transfer Bandwidth
[ 3] 0.0-10.0 sec 1.10 GBytes 943 Mbits/sec
root@bpi-r2:~# iperf -c 192.168.0.21
------------------------------------------------------------
Client connecting to 192.168.0.21, TCP port 5001
TCP window size: 43.8 KByte (default)
------------------------------------------------------------
[ 3] local 192.168.0.15 port 35150 connected with 192.168.0.21 port 5001
[ 102.243908] ------------[ cut here ]------------
[ 102.248532] WARNING: CPU: 0 PID: 0 at net/sched/sch_generic.c:461 dev_watchdog+0x278/0x27c
[ 102.256768] NETDEV WATCHDOG: eth1 (mtk_soc_eth): transmit queue 0 timed out
[ 102.263691] Modules linked in: mtk_thermal thermal_sys mt6577_auxadc spi_mt65xx pwm_mediatek mtk_pmic_keys nvmem_mtk_efuse ip_tabless
[ 102.276356] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 4.19.4-bpi-r2-testing #231
[ 102.283703] Hardware name: Mediatek Cortex-A7 (Device Tree)
[ 102.289259] [<c0114ca0>] (unwind_backtrace) from [<c010e9cc>] (show_stack+0x20/0x24)
[ 102.296947] [<c010e9cc>] (show_stack) from [<c0bbcff4>] (dump_stack+0x94/0xa8)
[ 102.304117] [<c0bbcff4>] (dump_stack) from [<c01283c0>] (__warn+0xf0/0x108)
[ 102.311027] [<c01283c0>] (__warn) from [<c0128430>] (warn_slowpath_fmt+0x58/0x74)
[ 102.318455] [<c0128430>] (warn_slowpath_fmt) from [<c096b95c>] (dev_watchdog+0x278/0x27c)
[ 102.326574] [<c096b95c>] (dev_watchdog) from [<c01a9008>] (call_timer_fn+0x4c/0x194)
[ 102.334261] [<c01a9008>] (call_timer_fn) from [<c01a9240>] (expire_timers+0xf0/0x16c)
[ 102.342031] [<c01a9240>] (expire_timers) from [<c01a95a8>] (run_timer_softirq+0xf4/0x1ec)
[ 102.350146] [<c01a95a8>] (run_timer_softirq) from [<c01023f8>] (__do_softirq+0x148/0x3b4)
[ 102.358265] [<c01023f8>] (__do_softirq) from [<c012f188>] (irq_exit+0xdc/0x11c)
[ 102.365521] [<c012f188>] (irq_exit) from [<c018b1f8>] (__handle_domain_irq+0x70/0xc4)
[ 102.373291] [<c018b1f8>] (__handle_domain_irq) from [<c0102268>] (gic_handle_irq+0x5c/0xa0)
[ 102.381577] [<c0102268>] (gic_handle_irq) from [<c0101a0c>] (__irq_svc+0x6c/0x90)
[ 102.388997] Exception stack(0xc1201ed8 to 0xc1201f20)
[ 102.394006] 1ec0: 00000000 00154614
[ 102.402121] 1ee0: df59e408 c0121760 ffffe000 c1204c70 c1204cb8 00000001 00000001 c12b8023
[ 102.410234] 1f00: c0e51ee4 c1201f34 c1201f38 c1201f28 c010a9f8 c010a9fc 60050013 ffffffff
[ 102.418351] [<c0101a0c>] (__irq_svc) from [<c010a9fc>] (arch_cpu_idle+0x48/0x4c)
[ 102.425695] [<c010a9fc>] (arch_cpu_idle) from [<c0bd8cd8>] (default_idle_call+0x30/0x3c)
[ 102.433726] [<c0bd8cd8>] (default_idle_call) from [<c015cafc>] (do_idle+0xf4/0x164)
[ 102.441326] [<c015cafc>] (do_idle) from [<c015ce2c>] (cpu_startup_entry+0x28/0x2c)
[ 102.448839] [<c015ce2c>] (cpu_startup_entry) from [<c0bd26ec>] (rest_init+0xc0/0xc4)
[ 102.456525] [<c0bd26ec>] (rest_init) from [<c1100ff8>] (start_kernel+0x46c/0x49c)
[ 102.463982] ---[ end trace 18e53bf3e74400e7 ]---
[ 102.468585] mtk_soc_eth 1b100000.ethernet eth1: transmit timed out
[ ID] Interval Transfer Bandwidth
[ 3] 0.0-10.0 sec 1.04 GBytes 897 Mbits/sec
happens after i made a iperf (client) over wan, set wan down, add ip to lan0 an make iperf again
root@bpi-r2:~# iperf -c 192.168.0.21
------------------------------------------------------------
Client connecting to 192.168.0.21, TCP port 5001
TCP window size: 43.8 KByte (default)
------------------------------------------------------------
[ 3] local 192.168.0.42 port 36400 connected with 192.168.0.21 port 5001
[ ID] Interval Transfer Bandwidth
[ 3] 0.0-10.0 sec 1.09 GBytes 940 Mbits/sec
root@bpi-r2:~# ifdown wan
[ 65.668446] mt7530 mdio-bus:00 wan: Link is Down
root@bpi-r2:~# ip addr add 192.168.0.15/24 dev lan0
root@bpi-r2:~# [ 86.884496] mt7530 mdio-bus:00 lan0: Link is Up - 1Gbps/Full - flow control off
[ 86.891791] IPv6: ADDRCONF(NETDEV_CHANGE): lan0: link becomes ready
root@bpi-r2:~#
root@bpi-r2:~# iperf -c 192.168.0.21
kernel:
root@bpi-r2:~# uname -a
Linux bpi-r2 4.19.4-bpi-r2-testing #231 SMP Tue Nov 27 18:58:30 CET 2018 armv7l GNU/Linux
reproducable with 4.19, 4.20 but not 4.14 so it seems i made i mistake while porting
sometimes it happens after trying wan again (mtk_soc_eth 1b100000.ethernet eth0: transmit timed out)…tried also with reverted in_atomic and clock-patch (no change)
it seems that some of the ethernet-Patches causes this crash and that they are not needed (as i thought). I applied the patch-series (cherry-picked from 4.20-gmac_test_dsa_only) i’ve posted to mainline-kernel to 4.19…
after testing with 4.20 and now applied to (new) 4.19-gmac and tested again…seems to work without problems
here my tests (iperf over wan,lan0,wan again and lan0 again):
log
root@bpi-r2:~# uname -r
4.19.10-bpi-r2-gmac
root@bpi-r2:~# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
valid_lft forever preferred_lft forever
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000
link/ether 9a:da:30:8e:aa:52 brd ff:ff:ff:ff:ff:ff
3: eth1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000
link/ether 5e:06:08:25:eb:ef brd ff:ff:ff:ff:ff:ff
4: wan@eth1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000
link/ether 5e:06:08:25:eb:ef brd ff:ff:ff:ff:ff:ff
inet 192.168.0.11/24 brd 192.168.0.255 scope global wan
valid_lft forever preferred_lft forever
5: lan0@eth0: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN group default qlen 1000
link/ether 9a:da:30:8e:aa:52 brd ff:ff:ff:ff:ff:ff
6: lan1@eth0: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN group default qlen 1000
link/ether 9a:da:30:8e:aa:52 brd ff:ff:ff:ff:ff:ff
7: lan2@eth0: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN group default qlen 1000
link/ether 9a:da:30:8e:aa:52 brd ff:ff:ff:ff:ff:ff
8: lan3@eth0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue state LOWERLAYERDOWN group default qlen 1000
link/ether 9a:da:30:8e:aa:52 brd ff:ff:ff:ff:ff:ff
9: wan.60@wan: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000
link/ether 5e:06:08:25:eb:ef brd ff:ff:ff:ff:ff:ff
inet 192.168.60.1/24 brd 192.168.60.255 scope global wan.60
valid_lft forever preferred_lft forever
10: lxcbr0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UNKNOWN group default qlen 1000
link/ether 8e:c6:30:ed:4c:43 brd ff:ff:ff:ff:ff:ff
inet 10.0.3.1/24 brd 10.0.3.255 scope global lxcbr0
valid_lft forever preferred_lft forever
root@bpi-r2:~# iperf -c 192.168.0.21
------------------------------------------------------------
Client connecting to 192.168.0.21, TCP port 5001
TCP window size: 43.8 KByte (default)
------------------------------------------------------------
[ 3] local 192.168.0.11 port 36498 connected with 192.168.0.21 port 5001
[ ID] Interval Transfer Bandwidth
[ 3] 0.0-10.0 sec 1.09 GBytes 940 Mbits/sec
root@bpi-r2:~# ip link set wan down
[ 129.232523] mt7530 mdio-bus:00 wan: Link is Down
root@bpi-r2:~# ip addr add 192.168.0.19/24 dev lan0
root@bpi-r2:~# ip link set lan0 up
[ 144.462925] mt7530 mdio-bus:00 lan0: configuring for phy/gmii link mode
root@bpi-r2:~# [ 150.717531] mt7530 mdio-bus:00 lan0: Link is Up - 1Gbps/Full - flow control off
root@bpi-r2:~#
root@bpi-r2:~# iperf -c 192.168.0.21
------------------------------------------------------------
Client connecting to 192.168.0.21, TCP port 5001
TCP window size: 43.8 KByte (default)
------------------------------------------------------------
[ 3] local 192.168.0.19 port 38498 connected with 192.168.0.21 port 5001
[ ID] Interval Transfer Bandwidth
[ 3] 0.0-10.0 sec 1.10 GBytes 942 Mbits/sec
root@bpi-r2:~# ip link set lan0 down
[ 233.152510] mt7530 mdio-bus:00 lan0: Link is Down
root@bpi-r2:~#
root@bpi-r2:~#
root@bpi-r2:~#
root@bpi-r2:~# ip link set wan up
[ 289.553227] mt7530 mdio-bus:00 wan: configuring for phy/gmii link mode
root@bpi-r2:~# [ 292.717522] mt7530 mdio-bus:00 wan: Link is Up - 1Gbps/Full - flow control off
root@bpi-r2:~# iperf -c 192.168.0.21
------------------------------------------------------------
Client connecting to 192.168.0.21, TCP port 5001
TCP window size: 43.8 KByte (default)
------------------------------------------------------------
[ 3] local 192.168.0.11 port 36502 connected with 192.168.0.21 port 5001
[ ID] Interval Transfer Bandwidth
[ 3] 0.0-10.0 sec 1.09 GBytes 940 Mbits/sec
root@bpi-r2:~# ip link set wan down
[ 318.692540] mt7530 mdio-bus:00 wan: Link is Down
root@bpi-r2:~#
root@bpi-r2:~# ip link set lan0 up
[ 352.022970] mt7530 mdio-bus:00 lan0: configuring for phy/gmii link mode
root@bpi-r2:~# [ 355.117526] mt7530 mdio-bus:00 lan0: Link is Up - 1Gbps/Full - flow control off
root@bpi-r2:~# iperf -c 192.168.0.21
------------------------------------------------------------
Client connecting to 192.168.0.21, TCP port 5001
TCP window size: 43.8 KByte (default)
------------------------------------------------------------
[ 3] local 192.168.0.19 port 38502 connected with 192.168.0.21 port 5001
[ ID] Interval Transfer Bandwidth
[ 3] 0.0-10.0 sec 1.10 GBytes 942 Mbits/sec
root@bpi-r2:~#
fixed crash if cpu-option is not set and want to post it to mainline, but mails got currently blocked on infradead (same as hdmi). contacted postmaster
if anybody here want to try it here is the actual patchset:
contacted andrew and florian how to get further to bring it mainline (they want to avoid dts-option and prefer bridge by user)
in my tests it lookes like this is caused by additional network-patches (qdma,bql,…). after removing them i did not seen this on 4.19 again (made multiple iperf rounds switching from gmac0 to 1 and back)…hoping phylink gets merged soon and go into next LTS. phylink works better than mainline-driver. i guess this is because of wrong gmac-setup (both need to set to same mode regardless gmac1 does not support trgmii)
Will try it but of course I will add persistent mac address for wan interface to dts file (I can’t get internet without it). As I understand it’s second gmac?
Yes, but without luck. It’s still lost network ports randomly (can see crash in dmesg if its yet restores network but it’s not always) and this make me crazy. Have tried 4.14 - no errors but very poor pptp vpn and I need these two (network and pptp) work. So just put device on the shelf and wait till someone release stable kernel
