On Tue, Feb 16, 2016 at 12:37:57AM +0000, fx IWATA NOBUO wrote: > > > > What limitation? Why change it? What's the problem here... > > > > > > In original code, number of devices to be able to attach is 8. > > > It's came from VHCI_NPORTS, the default is 8. > > > It can be changed with compile option -DVHCI_NPORTS The maximum number > > > is limited to USB_MAXCHILDREN (31). > > > > Why not just make it USB_MAXCHILDREN by default? > > I'm not clear about your question. > > My requirement is more device can be attached than 31 to connect > many IoT devices. That's an odd requirement to use the USB protocol for, especially when that is a hard-coded restriction in the USB protocol from what I recall. > > >From USB spec, USB_MAXCHILDREN is 31. > Max number of ports in a hub. > Defined include/uapi/linux/usb/ch11.h#L23 Yes, don't go over that. > If VHCI is modified to handle topology, it can be extends to 127 > including intermediate hub. > It need to introduce code to manage the topology but it's not done. > If it's done, the limit still be bound to 127. Again, a USB protocol limitation, something you really can't break. > So I tried more simple way by adding vhci controllers. > > Original USB/IP (before this patch): > Default is 8. Changeable by -DVHCI_NPORTS up to USB_MAXCHILDREN. > > This patch: > Default is 31. Changeable by -DVHCI_NPORTS up to USB_MAXCHILDREN. > Negative in previous e-mail, so I may change 8. > To reduce most outstanding overhead - event threads for each port, > I made patch 2/2. I don't understand at all, sorry. Please break your patches up, they should explain all of this much better than this email thread :) thanks, greg k-h -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
