On Thu, Jan 17, 2013 at 3:32 AM, Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx> wrote: > Also, why does this driver need something that the hundreds of other > serial drivers we have in-kernel today do not? What makes it special > over everything else? So (and I think Alan just seconded this) I think several vendors already have different ugly out-of-tree hacks to solve exactly this. The reason you haven't seen it before is that there is no AT command to power save a US Robotics 33k8 modem. And these are handled from userspace. The embedded systems have: - In-kernel driver using the TTY - Control over both ends of the link (we have the equivalent of an AT command setting the 33k8 into retention). - Critical (as in really critical) power requirements. Blueetooth chips on megabit serial links seem to be pretty common actually. Maybe people even have modems on these. I think the reason you haven't seen it before is that either people are scared of the TTY layer (whyyyy would they be that) or that they are in the out-of-tree business. We're trying to be in the in-tree business, so to be there we need to put the infrastructure in place to be able to be there. But I'm looking around for a *simple* UART-connected device with a low-power/retention state to use as example in question, so we don't have to conflate the CG2900 upstreaming issue with the TTY/UART PM issue. Anyone have hints on a suitable device? Linus Walleij -- To unsubscribe from this list: send the line "unsubscribe linux-serial" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html