On Wed, 26 May 2010, Matthew Garrett wrote: > On Wed, May 26, 2010 at 11:51:12AM -0400, Alan Stern wrote: > > (leaving the wakeup stuff to Rafael for the moment...) > > > Another question... The PCI core doesn't call pm_runtime_enable() or > > pci_runtime_set_active() anywhere. Apparently this is left up to the > > individual device driver. But shouldn't the core take this > > responsibility? What happens to PCI devices for which there is no > > driver? Shouldn't the core put them into runtime suspend by default? > > This is problematic in the GPU case, at least, since there are still > many that get driven by userspace rather than by the kernel. I don't know how that works. However an easy approach would be to make opening the device file (or however the userspace driver gets a reference to the device) cause the core to do a pm_runtime_get_sync(), with a corresponding pm_runtime_put_sync() when the file is closed. Alan Stern -- To unsubscribe from this list: send the line "unsubscribe linux-pci" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
