Why is this happening ? On different devices. on tl-wr942n Its also happening with eth connected to build in switch so it cant be temporal loss of signal in the wire.
[ 4713.290968] eth1: link down
[ 4717.291505] ar71xx: pll_reg 0xb8050014: 0x1099
[ 4717.295963] eth1: link up (100Mbps/Full duplex)
[ 4726.311027] eth1: link down
[ 4727.311466] ar71xx: pll_reg 0xb8050014: 0x1099
[ 4727.315920] eth1: link up (100Mbps/Full duplex)
[ 4747.354817] eth1: link down
[ 4748.355353] ar71xx: pll_reg 0xb8050014: 0x1099
[ 4748.359805] eth1: link up (100Mbps/Full duplex)
[ 4766.394749] eth1: link down
[ 4767.395291] ar71xx: pll_reg 0xb8050014: 0x1099
[ 4767.399744] eth1: link up (100Mbps/Full duplex)
[ 4813.422519] eth1: link down
[ 4814.423305] ar71xx: pll_reg 0xb8050014: 0x1099
[ 4814.427761] eth1: link up (100Mbps/Full duplex)
[ 4913.458074] eth1: link down
[ 4914.458594] ar71xx: pll_reg 0xb8050014: 0x1099
[ 4914.463051] eth1: link up (100Mbps/Full duplex)
[ 5075.484211] eth1: link down
[ 5076.484757] ar71xx: pll_reg 0xb8050014: 0x1099
[ 5076.489215] eth1: link up (100Mbps/Full duplex)
[ 5199.511533] eth1: link down
[ 5200.512040] ar71xx: pll_reg 0xb8050014: 0x1099
[ 5200.516495] eth1: link up (100Mbps/Full duplex)
[ 6770.548926] eth1: link down
[ 6771.549455] ar71xx: pll_reg 0xb8050014: 0x1099
[ 6771.553905] eth1: link up (100Mbps/Full duplex)
[ 8399.600048] eth1: link down
[ 8400.604549] ar71xx: pll_reg 0xb8050014: 0x1099
[ 8400.609002] eth1: link up (100Mbps/Full duplex)
[ 8408.631993] eth1: link down
[ 8409.632452] ar71xx: pll_reg 0xb8050014: 0x1099
[ 8409.636908] eth1: link up (100Mbps/Full duplex)
[ 8713.646164] eth1: link down
[ 8714.650507] ar71xx: pll_reg 0xb8050014: 0x1099
[ 8714.654956] eth1: link up (100Mbps/Full duplex)
Wed Nov 1 18:00:25 2017 kern.info kernel: [ 2176.590722] eth1: link down
Wed Nov 1 18:00:25 2017 kern.info kernel: [ 2176.594145] br-lan: port 1(eth1) entered disabled state
Wed Nov 1 18:00:25 2017 daemon.notice netifd: Network device 'eth1' link is down
Wed Nov 1 18:00:26 2017 kern.info kernel: [ 2178.093674] eth1: link up (1000Mbps/Full duplex)
Wed Nov 1 18:00:26 2017 kern.info kernel: [ 2178.100246] br-lan: port 1(eth1) entered forwarding state
Wed Nov 1 18:00:26 2017 kern.info kernel: [ 2178.105937] br-lan: port 1(eth1) entered forwarding state
Wed Nov 1 18:00:26 2017 daemon.notice netifd: Network device 'eth1' link is up
Wed Nov 1 18:00:28 2017 kern.info kernel: [ 2180.107833] br-lan: port 1(eth1) entered forwarding state
Wed Nov 1 18:00:55 2017 kern.info kernel: [ 2206.608588] eth1: link down
Wed Nov 1 18:00:55 2017 daemon.notice netifd: Network device 'eth1' link is down
Wed Nov 1 18:00:55 2017 kern.info kernel: [ 2206.612078] br-lan: port 1(eth1) entered disabled state
Wed Nov 1 18:00:56 2017 kern.info kernel: [ 2208.109374] eth1: link up (1000Mbps/Full duplex)
Wed Nov 1 18:00:56 2017 kern.info kernel: [ 2208.114456] br-lan: port 1(eth1) entered forwarding state
Wed Nov 1 18:00:56 2017 kern.info kernel: [ 2208.120172] br-lan: port 1(eth1) entered forwarding state
Wed Nov 1 18:00:56 2017 daemon.notice netifd: Network device 'eth1' link is up
Wed Nov 1 18:00:58 2017 kern.info kernel: [ 2210.123556] br-lan: port 1(eth1) entered forwarding state
From the log its clear that eth1 is br-lan. Its connected to internal switch.
Recently ag71xx driver was modified with GRO patch and I had severe problems from it on these 2 devices.
Now compiling version with that patch reverted. WIll see if it helps.
I won't be able to reflash tplink, but I can reflash d-link
You are right, eth1 is on internal switch according to log.
My tp-link 841n is still on openwrt CC. Maybe I can flash the latest LEDE trunk when get chance.
Found that link down/up triggers when the only PC connected to internal switch is shut down or rebooted.
Loss of connectivity on one of the switch ports causes eth1 to down/up for 1..5 seconds.
Is this normal behaviour ?
Its hard to test another combinations because this router is remote