Anyway, this isn't going anywhere. You offered to help, I told you what you can do to help (wait for rework to be done then test it). You have suggested various things, we have explained why they don't work or why we don't want to do them.
What it comes down to is this: I have chosen to do nothing to the code in svn for now, because I don't think it's a good idea to add an option which simply ignores all considerations of safety and spec compliance and leaves it in the users' hands. The other developers with whom I have discussed this at some length on IRC seem happy with this: we have been disabling Rockbox USB in release builds for iPods for several releases now, so leaving it disabled a little longer and fixing this properly isn't a big deal.
I don't think I have really explained (here, at least) exactly what I am doing and what I intend to commit, and maybe that's why you are continuing this rather fruitless thread, so here is my plan, which comes in two steps:
1) Rework USB charging so that the USB stack can instruct the hardware-specific code as to how much power is acceptable to draw at the present time, and add a third choice to the current USB charging setting so that it's a choice between "off", "on", and "force".
Off will ensure Rockbox draws at most 100mA (perhaps less if possible, though consideration of booting a device with a dead battery is needed). It will also instruct the USB stack to only request 100mA in the first place, which will allow connecting Rockbox players to bus-powered hubs (this is not currently possible, as we always request 500mA even when we don't use it).
On will allow Rockbox to draw 500mA when connected to an active host which has given us permission to do so. AC chargers, or anything else connected to the USB port, will still only be allowed to draw 100mA at this point. This will be the default setting, and will allow normal users to charge from a PC in the same way as with the OF.
Force will behave exactly the same as On when connected to an active host: 100mA until given permission to draw 500mA. However, if no host connection is detected within several seconds of detecting the physical presence of the USB connector, 500mA charging will be enabled anyway, for the benefit of AC chargers. It is on the users' head if they set this setting to Force and then connect it to something they shouldn't, but we will still do everything we can to prevent this causing a problem (which is why it will not override an uncooperative but active host).
2) Implement USB charger detection for ipodvideo and possibly other PP ipod models (though charger detection in the OF appears to work very inconsistently, other models may be hard). There are at least two ways to positively identify a charger, and from disassembling the OF I am reasonably sure the ipodvideo implements both: one is to connect an ADC to the USB D+ and D- pins and measure the voltages there; Apple chargers tie them to particular values which can be spotted. The other is to check if D+ and D- are shorted together, which is the USB 2.0 charging standard, and later Apple chargers also do this (the iPhone and iPod Touch appear to require this method). If any known method of positively identifying a charger succeeds, then full power charging will be enabled without requiring the Force setting. Some charger/device combinations may in fact be able to charge at higher than 500mA, and in the case of positive identification this will be allowed.
After 1) other developers with knowledge of other targets may be able to implement this USB charging functionality on different players, and may also be able to implement other kinds of charger detection on various players (e.g. mini/micro USB connectors have a fifth pin which can be tied to particular levels to signal presence of a charger).
I have done maybe half the work for 1) already, and I just haven't had time to finish it and test it on the players I have yet. When 1) is ready I will post a comment on FS#8802 so that interested people can test it; I'll even make you a test build if you can't build yourself
Does that make it clearer?
Topic: Investigating the iPod USB charging bug (Read 7195 times)
