On Thu, Aug 05, 2010 at 09:10:08AM -0500, Tabi Timur-B04825 wrote:
On Aug 5, 2010, at 8:52 AM, "Mark Brown" broonie@opensource.wolfsonmicro.com wrote:
Are you saying that every DT for a system doing audio on a PowerPC system has to manually specify the depth of the FIFO on the silicon?
Only for SSI devices.
Right, but that's the audio controller on the overwhelming majority of PowerPC devices.
That doesn't sound particularly sensible, and I had been under the impression that DT systems had a better way of coping with this.
The whole point behind having a device tree is to specify device information that cannot be probed. So every piece of information in the DT is something that the driver needs to be told because there is no way to query the hardware.
This seems crazy, it means that we're not able to use new support for hardware features to the driver which require any kind of flag or data without also going through and updating the device trees for all existing boards. That doesn't seem terribly helpful.
I'm fairly sure that some of the previous discussion with other device tree people suggested that this was something that there was infrastructure to cope with.