New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
FS#606 - Stall on Mediatek #5933
Comments
karesch: Similar issue reported also here: [[https://forum.lede-project.org/t/lede-v17-01-0-rc1/1285/11|External Link]], relayd, rp_filter is suspected there |
nbd: Please make a build with CONFIG_KERNEL_KALLSYMS enabled and post a fresh log |
karesch: Hi Felix, I would love to do that, but, unfortunately, making new builds is way beyond my abilities. I'm just a user. Sorry for that. |
rogerpueyo: Hi Karoly, I've compiled a sysupgrade image for you with CONFIG_KERNEL_KALLSYMS. Please find it at any of the following links: |
karesch: Hi Roger, You were quick, I'm impressed. I'll do the testing this evening and will come back to you. Thanks a lot, |
karesch: This is the 1st log, I'll submit further bugs: |
karesch:
|
karesch: under heavy load, the system crashes and reboots |
karesch: it's still the same. I have the skbuff problem, what you can see above, just 1-2 minutes after boot. Under load (iperf3 or heavy torrent) the router crashes and reboots, I have no chance to look at the DMESG or syslog. |
bjonglez: Karoly, you can try looking at /sys/kernel/debug/crashlog after a crash-reboot, it might hold an interesting stack trace. |
karesch: Hi Baptiste, Thanks for the tip, I didn't know that. I checked the directory. Unfortunately, there was no //crashlog// file in the ///sys/kernel/debug/// directory. I waited for 3 crashes, no results. I also did //echo c >/proc/sysrq-trigger// to check, whether crashlog settings are OK. This method generated a crash file, but the bug unfortunately, not. Do you have any other idea to get a crashlog file? Thanks, |
karesch: Hi Baptiste, I could copy-paste this: |
karesch:
|
karesch: there is still no crashlog, the above test was done on the version provided (//LEDE Reboot 17.01-SNAPSHOT r3276-4a405ac8f9 / LuCI lede-17.01 branch (git-17.063.59066-a5191ef)//) with SQM cake switched on on WAN. I attached the dmesg output. |
karesch: Another interesting symptom was that on snapshot r3276 the ac wifi speed was smaller lower than on 17.01 release (80-100 Mbit/s versus 220-300 Mbit/s) |
karesch: a few minutes after the above stalls the rooter rebooted, no crashlog found in /sys/kernel/debug/ directory: |
bjonglez: This has been fixed, see FS#804 |
karesch:
On D-Link DIR-860L B1, processor MediaTek MT7621
LEDE 17.01.0
The system used connected to WAN and wifi is used (ac and n). WDS is also used (as server) and SQM is used on WAN. Only the WAN switch port is used.
After random time wifi is disconnected and connected again, can be seen in the log as: WARNING: CPU: 1 PID: 0 at net/core/skbuff.c:4194. However, after some time, which can be minutes or days wifi totally disconnects, and even DHCP does nor work on the LAN ports. The log shows: INFO: rcu_sched self-detected stall on CPU. Only a reboot can help. Sometimes the device reboots by itself.
The above could not be linked to any event (wifi connection, change in sqm settings, etc).
Switched back to Openwrt 15.05.1.
A similar issue was reported here: [[https://forum.lede-project.org/t/build-for-the-d-link-dir-860l/948|External Link]]
The text was updated successfully, but these errors were encountered: