northbound:

I am not sure if this is a clue or not.
4.4 runs about 340 bits entropy and 4.9 = 3.3k bits.
This is with identical settings except for kernel.

lister-wrt:

@jeff that definitely sounds like something that could be an issue for us. Any idea what a healthy amount of entropy is?

northbound:

I honestly do not know what would be considered a good entropy bitrate..Sorry.
I still think it comes down to a timing issue or below.
Sat Jul 15 17:30:04 2017 kern.err kernel: [ 1.276889] cpu cpu1: opp_list_debug_create_link: Failed to create link
Sat Jul 15 17:30:04 2017 kern.err kernel: [ 1.283536] cpu cpu1: _add_opp_dev: Failed to register opp debugfs (-12)
I have been up from kernel bump to kernel bump for awhile now. No reboots for days at a time but when it happens it sure would be helpful to be able to grab a crash dump.
I do not know if the new GCC version has anything to do with it. Up time that is.
[ 0.000000] Linux version 4.9.38 (jeff@jeff-VM) (gcc version 7.1.0 (LEDE GCC 7.1.0 r4209-62d0b1a) ) #0 SMP Sat Jul 15 21:13:34 2017
r4583-e4e984f
I do not even want to think about the hours I have spent to find out what the opp problem is.
It is beyond me.

Hopefully someone that knows what they are doing will get to the bottom of the issue.

I have also been using diizzyy's pull req. lede-project/source#1211

lister-wrt:

InkblotAdmirer had something to say about this issue today;

https://forum.lede-project.org/t/wrt1900acv1-reboots-kernel-4-9/2025/62

It's way beyond me but hopefully this is useful information to anyone reading this :)

lister-wrt:

Issue persists. Others have expressed interest in solving. See discussion.

nbd:

Please try the latest version from my staging tree at https://git.lede-project.org/?p=lede/nbd/staging.git;a=summary
I’ve pushed an IRQ related fix, maybe it will help with the stability issues.

northbound:

I am running the latest 4.9.58 using your patch 4.9/130-irqchip-armada-xp-backport.patch
Will let you know how it goes

northbound:

No crashlog and a reboot in about 2 hrs will bump kernel now

nbd:

Fix pushed to master in r5355-31691f9649

NainKult:

Asking to reopen the issue. Users are still experiencing random reboots on mamba. https://forum.lede-project.org/t/wrt1900acv1-reboots-kernel-4-9/2025/91

We should be able to identify the issue by solving and closing [[https://bugs.lede-project.org/index.php?do=details&task_id=564|#564]]

nbd:

From what I see in the forum, it seems that the main crashes have been fixed and the remaining ones are an mwlwifi specific issue which is probably not limited to mamba.
Or am I missing something?

NainKult:

The randomness of the event and the lack of backtrace makes it hard to pinpoint.

However, a crash occurred on my unit, radios off and kernel module unloaded. Seems that mwlwifi issue is unrelated.

Edit 1: You may want to wait for the mwlwifi issue to be resolved first to make sure. I would do the same if I were you.