RMerlin, how do you test your firmware?

[Kobold]

Dear RMerlin, do you use anything special to test your firmware? Some kind of a MIPS virtual machine perhaps?
Is it possible to test a firmware for ASUS RT-AC66U without actual flashing?
Thank you.

 

[RMerlin]

There's no emulator, since you'd need to emulate all the SoCs and wireless HW in addition to the CPU. So testing usually involves a lot of flashing. Thankfully I have a spare RT-N66U for that, while the RT-AC66U can stay up and running as the main router.

Occasionally I can test just a modified binary by mount binding it on top of the flashed one. This is mostly the case for httpd-specific changes, since it can easily be restarted at any point.

 

[netmik3]

Would love to see you on youtube

 

[RMerlin]

Nothing interesting to see there really

But here's a pair of shot of tonight's test setup, which is probably more interesting:

Wish I had kept a picture of the setup I had back when I was working on OpenVPN support. My living room had:

- One RT-AC66U (my main router, providing WAN and LAN)
- One RT-N66U (used as an OpenVPN client)
- One WRT310N running DD-WRT (used as an OpenVPN server)
- My laptop (connected to both the RT-N66U as a client, and the RT-AC66U for access to my development VM)
- My eeePC (target machine at the other end of the tunnel)

 

[netmik3]

beautiful

 

[ryzhov_al]

Oh, you soldered serial console to catch NFS kernel panic. Good luck!

 

[RMerlin]

Oh, you soldered serial console to catch NFS kernel panic. Good luck!

Nothing soldered - there are pin headers on the router. Just some jump cables that came with my serial adapters

I like the fact they have a built-in FTDI chip, so I can plug them directly to a USB port (using an extension cable here for convenience).

No luck in catching the cause of the NFS issues however. I applied a few missing NFS patches, also applied a few additional patches that might be related, still no luck.

I opened an issue on Github, in case you (or anyone else) feels like digging more in depth.

Are you positive the issue does not exist with neither Tomato or WL500G? There are so many patches to the WL500G kernel that it's like trying to find a needle in a haystack figuring out which patch fixed this issue. I can't find any reference to the bug itself or to anything specific to the kernel function where the crash occurs.

Tomato's kernel has fewer patches, so it might be easier to track it down, assuming the patch appears in their Git logs.

 

[ryzhov_al]

Are you positive the issue does not exist with neither Tomato or WL500G?

wl500g is definitely free of this issue.

 

[lucian]

does this bug apply only as client or also as nfs server?