|
|
|
|
The current implementation allows the user to specify the desired CPU
|
|
|
|
|
clock value, in MHz, via an environment variable "cpuclk".
|
|
|
|
|
|
|
|
|
|
Four compile-time constants are used:
|
|
|
|
|
|
|
|
|
|
CONFIG_8xx_OSCLK - input quartz clock
|
|
|
|
|
CONFIG_SYS_8xx_CPUCLK_MIN - minimum allowed CPU clock
|
|
|
|
|
CONFIG_SYS_8xx_CPUCLK_MAX - maximum allowed CPU clock
|
|
|
|
|
CONFIG_8xx_CPUCLK_DEFAULT - default CPU clock value
|
|
|
|
|
|
|
|
|
|
If the "cpuclk" environment variable value is within the CPUCLK_MIN /
|
|
|
|
|
CPUCLK_MAX limits, the specified value is used. Otherwise, the
|
|
|
|
|
default CPU clock value is set.
|
|
|
|
|
|
|
|
|
|
Please make sure you understand what you are doing, and understand
|
|
|
|
|
the restrictions of your hardware (board, processor). For example,
|
|
|
|
|
ethernet will stop working for CPU clock frequencies below 25 MHz.
|
|
|
|
|
|
|
|
|
|
Please note that the new clock-handling code is enabled if
|
|
|
|
|
CONFIG_8xx_CPUCLK_DEFAULT is defined. Since this mechanism supports
|
|
|
|
|
only MPC866 and newer CPUs, this constant MUST NOT be defined for
|
|
|
|
|
MPC823/850/860/862 series. The clock generation algorithm for older
|
|
|
|
|
chips is different and has not been implemented yet. If you need it,
|
|
|
|
|
your patch is welcome.
|
