I'm not saying that the kernel shouldn't initialize the attributes or have a default. But it should only set the default when the attribute is initialized (It doesn't even matter to me whether it's enabled or disabled). It's just there should not be further manipulation from the kernel (e.g. device_set_wakeup_enable) afterwards because 1. it's brings inconsistency because the function is adopted per driver 2. it's a user preference and responsibilty 3. third it prevent udev to apply a rule properly (regression / bug) P.S. Alan for my case, I don't need a patch for logitech-dj, I just need to remove device_set_enable_wakeup from hid_core.c, then I can enable or disable the attribute with a udev rule happily for both devices. ▶ Show quoted text On 23 April 2015 at 09:21, Peter Chen <Peter.Chen@xxxxxxxxxxxxx> wrote: > >> > > >> > > Oh, okay, I didn't realize that. >> > > >> > > Is there a reasonable way to enable wakeup only when the driver >> > > learns that a keyboard is connected? Where would the driver do this? >> > >> > I don't know if the driver ever "knows" this, as you can pair lots of >> > different devices to this same receiver. There's a userspace >> > application that lets you manage the device, called "solaar", that >> > this option could be changed in. >> > >> > But really, putting the device to sleep should work for it no matter >> > if this is a keyboard or a mouse or a joystick, as the wakeup logic is >> > in the receiver, not in the device on the other end of the wireless link. >> > >> > Turning autosuspend works for me for my mouse connected. It doesn't >> > work for one of my "real" USB keyboards when it's connected to the >> > machine, which is why I can't enable autosuspend for it, as it drives >> > me crazy. >> > >> > I don't have a keyboard to test the receiver with at the moment, to >> > see if autosuspend works for both things connected at the same time, >> > or for just the keyboard. >> >> Tom and I have been talking about enabling wakeup, not autosuspend. >> The question is whether or not the default wakeup setting for the receiver >> should be "enabled". >> >> The kernel's policy is that keyboards should be enabled for wakeup, by default. >> I think that matches most people's expectations. But when you've got a >> "universal" receiver, what then? >> >> Should it always be enabled by default because a keyboard _might_ be >> connected? Should it be enabled only when a keyboard is detected? >> What if multiple devices are connected at the same time? >> > > From my point, the user option should not depend on kernel default value. > If the system you build needs some USB devices as system wakeup source, > the developers need to make sure the wakeups are enabled before system > enters suspend. > > Peter > >> Shucks -- does the receiver even _work_ as a wakeup device? It claims to, but >> that would require it to remain in wireless contact with the remote keyboard >> even while it's supposed to be in a low-power state, which rather seems to >> defeat the purpose. >> >> Alan Stern >> >> PS: I've got wakeup enabled for the PS/2 keyboard attached to the computer >> I'm using now, but it doesn't work. I have to press the power button to wake >> the machine up from suspend. Probably an issue in the BIOS or ACPI -- I >> haven't bothered to try and track it down. >> >> -- >> 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 -- 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
