Wolfgang M�es wrote: > Hello, > > On Monday 30 September 2002 16:24, Alan Stern wrote: > >>Actually, it is possible to receive EPIPE for a control endpoint. But it >>shouldn't happen very often. In fact, with normal use we shouldn't see it >>at all. > > > Oh yes, I have seen it. EPIPE on control endpoints. Mostly with the usb-uhci > driver (Linux 2.4.10). It was no STALL but a transmission error. There was a > spike on the USB data lines, and the host is doing 3 retries. All 3 retries > failed because the spike was longer than all the retries.
So it seems like receiving -EPIPE is legal and common enough to notice in robust drivers for reasonable devices, both because of protocol stalls and transmission errors. (It'd be nice if those two were fully distinct, but ...) What was a bit surprising to me was the bit Alan quoted from [9.4.5] that said devices can actually set a halt condition on control endpoints for "functional error conditions" (though it's not recommended) if, as it says in [8.5.3.4], they're not "well designed" devices. I think that means that some drivers, maybe even usb-storage, might need to be able to clear halts on the control pipe, if they talk to devices that are not "well designed". BUT any driver that tries should expect that most devices return protocol stalls (-EPIPE) when asked to clear halts. Matt, does this help resolve your issue? - Dave ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
