On Fri, Mar 02, 2018 at 09:39:34PM +0000, Xianwen Chen wrote: > Dear Stefan, > > Thank you. I guess the first step is to find out a way to ask iwi0 > driver to output more details of the error, before it could be fixed. > I don't have prior experiences with the iwi0 driver. Do you know how I > can enable a debugger mode, if there is one? > > Yours sincerely, > Xianwen
You would have to edit the source code. There is no debugging mode. There is nobody willing to work on this, so you'd have to fix it yourself. I'd be happy to review and test patches. However, this won't be easy at all. The driver has almost no control over this hardware. Most things are handled inside the firmware and we cannot look inside. The driver only sees a single bit which says "a fatal error occured" with no further information available. The driver tries to recover from such errors automatically by resetting the device. There's nothing else it can do if such an error is reported. The only way to fix the problem would be to prevent the situation where the firmware error occurs, but we don't even know why it is happening. If the problem turned out to be a firmware bug it could not even be fixed in the firmware because nobody at Intel will care about this anymore :( There is much nicer hardware with better drivers and firmware available, both for regular usage and for hacking/debugging/learning. If you can open up the machine and swap out the miniPCI wifi card, an ath(4), old athn(4), or ral(4) card should work much better. These are similarly old devices and can be bought second hand. See the man pages for device names which can be searched on the web. Or you could use a USB device. I would recommend an AR9271 USB athn(4) device because it uses open source firmware, or a run(4) or urtwn(4) device.
