Comment 78 for bug 362875

I tried to do an apport and it seg-faulted in the middle of it. However a few days of researching and looking at the chipset datasheet specific to my model and I found something useful.

The dmesg report complains about vendor request 0x07 which is a multiple byte read in the specs, while the offset 0x3000 corresponds to the MAC control register, specifically the ASIC version number.

As well, the fact that this bug has existed across releases, kernel versions, and the transition to systemd points more to the driver than anything else. I also tested wicd (currently using that now) and the problem persists, which means network-manager isn't the primary source of the bug. All of these can still be in play or making the problem worse but aren't the sole cause.

If I can get sight into the drivers (both rt73 and rt2x00USB), I might be able to pin down more specifics. For the actual bug fixers, I'd advise to check and see if the non-USB version of this chipset shares the problem to isolate the bug some more.