On Monday 21 May 2007 13:22, Henrique de Moraes Holschuh wrote: > On Mon, 21 May 2007, Dmitry Torokhov wrote: > > RFkill is not only about "please disable radios", it _does_ provide > > the standard sysfs interface to control the radios as well. > > To control? Fine, if it was really controlling them. The problem is that > thinkpad-acpi needs it to *interface* with the firmware that really controls > the radios: it is not like thinkpad-acpi or the kernel is in *control* of > the radios. Ok, it interfaces the firmware that controls the radio in case of thinkpad_acpi. Is that precise enough? > > My issue with rfkill is that if the user queries the status of the rfkill > switch (for what thinkpad-acpi would need it for), he should get the status > of the rfkill switch in hardware, and not of whatever the kernel thinks it > is. > > > You know I got curious and looked at thinkpad_acpi driver and adding > > rfkill support for bluetooth to it turned out to be pretty easy. What > > It is easy, indeed... until the user slides the hardware rfkill switch to > "kill it". Now the kernel thinks the thing is on, and it has died. Etc. > And *nothing* you do can get the radio back on, unless the user slides the > hardware switch back to the on position, at which point it is anyone's guess > if the firmware in a given thinkpad will turn it ON, or leave it off if it > was off in software... > And once you implement a polling input device for your slider that should work too. I know you don't like polling. But what you seem to be missing is that you still need to notify userspace that your radio state is changed so you need to poll anyway. -- Dmitry ------------------------------------------------------------------------- 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
