On Mon, 14 May 2012, Matthieu CASTET wrote:
> >> + else {
> >> + reg = &ehci->regs->port_status[port];
> >> + bit = HOSTPC_PHCD << 1;
> >> }
> >> +
> >> + temp = ehci_readl(ehci, reg);
> >> + ehci_writel(ehci, temp & ~bit, reg);
> >
> > This can't be right. In the !has_hostpc case, you do not turn off the
> > RWC bits before writing them back.
> >
> I don't understand : the bit is turned of by "temp & ~bit" that is in the common
> path.
Sure, you turned off the HOSTPC_PHCD bit. But what about the PORT_CSC,
PORT_PEC, and PORT_OCC bits (collectively known as PORT_RWC_BITS)? If
any of them happened to be on when the ehci_readl() executed, the
ehci_writel() would turn them off. That's not what you want.
Alan Stern
--
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
