384.7 IPV6 Stateful Generates Logged Errors, Stateless Does Not

[themiron]

@GHammer

I don't even know if they are finished yet, but since I've been doing test builds I did a 'test2' build with the patch from the dnsmasq author replacing @theMIROn 's patch. Same download location.

7b885328c (HEAD -> test) dnsmasq: (upstream) properly deal with unaligned addresses in DHCPv6 packets
c3617d4d3 dnsmasq: (upstream) fix broken DNSSEC records in ef267f514
ef267f514 dnsmasq: (upstream) don't return NXDOMAIN to empty non-terminals
7223b63e5 Revert "dnsmasq: dhcpv6: fix unaligned access crash on aarch64"

Simon has decided to realign input data rather to support unaligned although more changes were required/size consumed.
Both aproaches has same result, bug is gone, at least for current sources.
Gonna to reuse upstream version a bit later, a bit more cosmetic changes might be expected.
FYI http://lists.thekelleys.org.uk/pipermail/dnsmasq-discuss/2018q3/012445.html

 

[GHammer]

I have had nearly a full day on test2 with nothing logged of any sort.

 

[john9527]

@GHammer
Thanks for the feedback. Good to hear.

Regarding your 'powerdown' when rebooting after loading the firmware, I don't think that's related at all to this change....it's been reported before.

@theMIROn - Your next mission should you decide to accept it The fork() function on the AC86 is randomly failing with an out of memory message, even with plenty of memory available. I think the random 'powerdown' on boot may be related as the reboot function uses fork() as part of it's error handling.

 

[GHammer]

Thanks for the feedback. Good to hear.

Regarding your 'powerdown' when rebooting after loading the firmware, I don't think that's related at all to this change....it's been reported before.

No, I don't think it is either. I've seen that before when I was running a scheduled restart.

 

[themiron]

The fork() function on the AC86 is randomly failing with an out of memory message, even with plenty of memory available.

no idea about, pre-3.19 mips kernels have fork issues, but I've never heard about something similar on 4.x arm/aarch64.
any steps to reproduce?

 

[GHammer]

no idea about, pre-3.19 mips kernels have fork issues, but I've never heard about something similar on 4.x arm/aarch64.
any steps to reproduce?

I'll defer to @john9527 but if I set a scheduled reboot, within 2 days I will have to use the power button to turn the router on. It simply powers off with no logs left behind.

 

[john9527]

no idea about, pre-3.19 mips kernels have fork issues, but I've never heard about something similar on 4.x arm/aarch64.
any steps to reproduce?

I think the original problem showed up in Skynet. @Adamm can you help out (or am I having a memory fault again )

 

[RMerlin]

no idea about, pre-3.19 mips kernels have fork issues, but I've never heard about something similar on 4.x arm/aarch64.
any steps to reproduce?

I think he's referring to the issue that is specific to HND, were fork() can randomly complain about running out of memory. When I talked with Asus about it, they told me it was a bug within the SDK, that only Broadcom could fix, but that they were going to implement a workaround. I couldn't get any further details out of them at the time.

Personally, I suspect the issue might lie in glibc, since that's one thing that's different in HND (older SDKs using uclibc).

 

[RMerlin]

I'll defer to @john9527 but if I set a scheduled reboot, within 2 days I will have to use the power button to turn the router on. It simply powers off with no logs left behind.

Probably a completely different issue. The RT-AC68U was (for some unknown reason) very prone to this as well. The issue is that the router crashes during the reboot, requiring a power cycle. That issue has been there for years, and is completely random (I have never been able to reproduce it myself).

 

[GHammer]

That issue has been there for years, and is completely random (I have never been able to reproduce it myself).

I've had great (?) luck reproducing it on my 86U. In fact, I RMA'd the first one I had because of this and then the second does exactly the same. Not really a huge issue for me because I do not have large amounts of traffic, connections, or addon programs/services.
Ir was annoying when the second one did this as I had left on vacation the very day it decided to power off. No weather data reported out and my servers at home were unreachable.
For me, it is easy. Simply schedule a reboot daily at 05:20 and wait 2 days for the dark router.