On Tue, 22 May 2007, Richard Hughes wrote: > > > Ick indeed. It should really be automatic (ie. no need to poke anywhere > > > to get the events back to user-space). > > > > No way, no how. The firmware often does better at it. > > Does better than userspace? By taking control of the backlight Yes, as in "it works". Userspace/linux doesn't know how to do all that the firmware does in every thinkpad out there, and I doubt it ever will, as that would require IBM and Lenovo to cooperate opening the firmware docs, etc. Setting that hotkey mask is model-specific knowledge. While I wouldn't mind much doing it in kernel space, I have been told to not do that. > Like I said before, on my X60, all the buttons do the hardware action > even with the hotkeys enable. I'll update the bios later today, and see > if this is still the case with the latest bios. So far it looks like the X60 changed the interface but kept non-functioning bits of the older interface around. Argh. > Well, I really wanted new keys in input.h, but we do need to get this > working out of the box. The number of users that are going to enable the > hotkeys at boot is going to be a fraction of one percent. I think we > need a way to (at least) to toggle this as a module option. Would you > accept a patch for this? No. That's why I am making the keycode map configurable. We have three choices, in order of preference: 1. New keys get accepted in input.h. 2. Someone starts writing a thinkpad userland helper to configure thinkpad-acpi based on DMI. I may take on this, if I have the time. But if someone offers to do so, I would like that a lot. This would be a good idea even if we get (1) above. 3. HAL or something else GNOME/KDE takes care of the keycode map, and I add a lock so that the admin can still override them and have it stick. Given how nosy KDE and GNOME are already becoming on thinkpads, to the point that I always ask people debugging thinkpad-acpi to do stuff at single user mode, I am very heavily biased for (1) AND (2) happening. BTW, currently I am extremely unamused on how incomplete and messed up sysfs is. Binary attributes are so much of a second-class-citizen, as are subdirectories, that I might have to change the thinkpad-acpi sysfs interface. It is a *good* think I warned in the docs the sysfs interface was still heavily experimental and not stable, so I will not accept any complains over it :-) -- "One disk to rule them all, One disk to find them. One disk to bring them all and in the darkness grind them. In the Land of Redmond where the shadows lie." -- The Silicon Valley Tarot Henrique Holschuh ------------------------------------------------------------------------- This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/ _______________________________________________ ibm-acpi-devel mailing list ibm-acpi-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.sourceforge.net/lists/listinfo/ibm-acpi-devel
