You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'll try to keep this short...
I have two PC's each with a "Aquantia Corp. AQC107" PCI-E card and direct eth connection between the two (no switches, etc).
One PC runs Windows 10 x64 (desktop), the other Debian x64 bookworm/sid on the 6.0.0 debian kernel (server).
The server is up 24/7, the deskop is on/off every day.
driver: atlantic
version: 6.0.0-6-amd64
firmware-version: 3.1.58
expansion-rom-version:
bus-info: 0000:02:00.0
supports-statistics: yes
supports-test: no
supports-eeprom-access: no
supports-register-dump: yes
supports-priv-flags: yes
Occasionally (say 25% of the time), the Linux server end of the Ethernet link is not reestablished after booting the Windows PC.
The only way ive found to re-establish the connection (without re-booting the Linux server) is to run the following script:
brctl delif br0 enp2s0
sleep 5s
ip link set enp2s0 down
sleep 20s
rmmod atlantic
sleep 15s
modprobe atlantic
sleep 15s
ip link set enp2s0 up
sleep 10s
brctl addif br0 enp2s0
brctl show
When I do so, most of the time the eth link then becomes operational.
Sometimes when I do run the script, i get segfaults in the kernel log, and the link does not come up.
It's pretty random.
Below is a section of the kernel log over a few days showing the link going up and down successfully, a fail, the segfaults that sometimes occur when running that script, and then the link coming up ok again.
Please let me know if i can provide an information that would help diagnose the issue.
linkup failure between the nodes. That could be cable or autoneg issues. Hard to say at the moment.
Panic in linux. That seems to come from out of memory condition. Could be a mem leak.
Could you try running in your system "ifconfig down enp2s0; ifconfig up enp2s0" in a loop? If memory leak, you should get the same issue immediately.
Please also try installing the driver from this repository and check the same.
I'll try to keep this short...
I have two PC's each with a "Aquantia Corp. AQC107" PCI-E card and direct eth connection between the two (no switches, etc).
One PC runs Windows 10 x64 (desktop), the other Debian x64 bookworm/sid on the 6.0.0 debian kernel (server).
The server is up 24/7, the deskop is on/off every day.
Occasionally (say 25% of the time), the Linux server end of the Ethernet link is not reestablished after booting the Windows PC.
The only way ive found to re-establish the connection (without re-booting the Linux server) is to run the following script:
When I do so, most of the time the eth link then becomes operational.
Sometimes when I do run the script, i get segfaults in the kernel log, and the link does not come up.
It's pretty random.
Below is a section of the kernel log over a few days showing the link going up and down successfully, a fail, the segfaults that sometimes occur when running that script, and then the link coming up ok again.
Please let me know if i can provide an information that would help diagnose the issue.
The text was updated successfully, but these errors were encountered: