> > > static void uhci_reset(void *opaque) > > > @@ -1240,8 +1239,7 @@ static int usb_uhci_common_initfn(PCIDevice *dev) > > > /* TODO: reset value should be 0. */ > > > pci_conf[USB_SBRN] = USB_RELEASE_1; // release number > > > > > > - s->irq_pin = u->info.irq_pin; > > > - pci_config_set_interrupt_pin(pci_conf, s->irq_pin + 1); > > > + pci_config_set_interrupt_pin(pci_conf, u->info.irq_pin + 1); > > > > So everyone does this + 1/ - 1 logic. > > > > We get comments like: > > pci_config_set_interrupt_pin(pci_conf, 1); /* interrupt pin 0 */ > > which just shows the API is confusing. > > How about we change pci_config_set_interrupt_pin to do + 1 > > internally? > > Then add pci_config_get_interrupt_pin to do - 1. It seems a good idea. > > > > Add a comment that this does not support devices without interrupts, > > for that - use get_byte directly. > > > Note: this is not a problem with your patch, it can > be a cleanup on top. > I will make another patch(not part of this series) to be applied on top of this series.
Thanks, Marcel
