Porting crashlog for kernel 5.4

Hi folks,

since I experiencing crashes on my TP-Link C2600, I need crashlog to get some insights why the kernel crashes.
I, therefore, porting the crashlog patch to Linux kernel 5.4 because the default kernel has changed from 4.19.X to 5.4.35.
Are there other people out there who trying the same?
Or is a developer around who has some knowledge about the crashlog patch and might wanna help me?


Getting random reboots (weekly, if not more often) with my C2600 also, and dumping logs to USB flash drive didn't reveal anything. I'm greeted with No memory allocated for crashlog so I'm looking for a way to enable this. Can you point me to the patch your're mentioning?


You can use the already available ramoops. For that, you have to manually edit the kernel config (make kernel_menuconfig). Activate the pstore filesystem (Misc Filesystems) and enable the option Log panic/oops to a RAM buffer.
Afterwards, change the default cmdline string in boot options for the kernel to mem=384M ramoops.mem_address=0x5fba4000 ramoops.mem_size=0x45bfff
After experiencing a crash-based reboot, mount the pstore pseudo-fs mount -t pstore pstore /mnt/. The files located /mnt contain your crash log.

1 Like

Thanks - I was looking for something similar!
In my case I was in fact still on 4.4 kernel (19.07.2 release) that had the crashlog patch but as far as I can see that doesn't work on IPQ8064. I've flashed current snapshot with 5.4 kernel and I'll use your instructions in the case I experience crashes. 5.4 didn't crash for me in about 24h, but I'm seeing rather high CPU usage from kernel itself (kworker/1:0+events hovers around 21%)

I've just reported my crash: https://bugs.openwrt.org/index.php?do=details&task_id=3099
If you encounter a similar problem, pls report your config, and +1 via the comment function.

After 5+ uneventful days I've noticed that my router has rebooted, so I've managed to follow your instructions and build an image with ramoops enabled kernel. I'll wait for the next reboot and see if it's due to an kernel oops and if it is I'll check if its more/less the same oops you caught. I took a peek at the kernel code involved and saw nothing that I can point my finger to - let's just hope we are chasing the same bug and go from there.

I saw that my router rebooted couple hours ago - here is my crashlog: https://pastebin.com/raw/fhD5vVLQ. It's obviously different than yours but I'm not really sure if this is the "original" crahslog or router rebooted more than once and I'm seeing the last one.

1 Like

Router rebooted again and crashlog is somehow exactly the same so I'm guessing ramoops doesn't really work 100% reliable for me. I'll plug the router out of the power (to clear the RAM) and plug it back in and wait for the next crash.

This topic was automatically closed 10 days after the last reply. New replies are no longer allowed.