On Wed, 2004-05-05 at 16:12, Eric W. Biederman wrote:
Li-Ta Lo ollie@lanl.gov writes:
On Wed, 2004-05-05 at 10:51, YhLu wrote:
If I changed config.lb PCI 1:0.2 on --> PCI 1:0.2 off The PCI 1:0.0 and PCI 1:0.1 will be disabled together.
According to AMD 8111 data sheet, LPC 0x48, only can be disabled the device in second bus by device num, and the three USB share one device number.
So I guess the code about 0x47 must be there.
Ron and I just dicurssed about the enable/disable mechanism for amd8111. We are going to implement a slightly better one soon.
Actually, the enable/disable for devices in amd8111 is multiplexed in amd8111_enable and the device_operation::enable() is directed to amd8111_enable. We are going to demultiplex this and each devices has its own "real" enable method.
No. This must be per chip.
Because you cannot necessarily see these devices before you enable/disable
What do you mean ? Do you mean when I try to enable/disable something on 8111 but the LPC bridge is not there yet ?
them. And only one device on the 8111 actually does the enable/disable.
The function amd8111_enable() is used both as .enable and .enable_dev. Why are you doing this ? This make it very difficult know under which contex the function is called and what it is supposed to do when it acts as different method.
What I want to do is make amd8111_enable() a generic public function and each device's .enable() calls amd8111_enable() to do the real thing.
Ollie