Kristian> Please read again what you wrote and think about it:
Kristian> Jes Degn Soerensen said:
Kristian> Seems to me that a) the current kernel interface is not
Kristian> complete enough for the approach (do as much as possible in
Kristian> userland) you propose and that b) you are proposing just
Kristian> what you say the KGI part of GGI does (extending the kernel
Kristian> interface), but with the additional disadvantage of doing it
Kristian> in an uncoordinated, random way.
We disagree strongly when it comes to putting things related to
acceleration into the kernel. I think this should only be in the
kernel for console support for frame buffer consoles and nothing else.
Kristian> The GGI people and you agree in how things should be done
Kristian> (do as much as possible in userland), and you both agree
Kristian> that kernel extensions are necessary to enable these
Kristian> userland application to do their work. But their proposal
Kristian> organizes the kernel modifications in a more meaningful
Kristian> way. Plus they already have the needed userspace emulation
Kristian> libraries ready.
The multihead support is not unstructured as it is now for monitor, a
structured way of assigning input devices to consoles does not have to
do with graphics at all. There may be a limitation in the way the
kernel handles this at the moment.
If people need interrupts for graphics boards, then there is a problem
which is currently not handled by the kernel.
Jes
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu