Lou, make sure you don't have a mismatch between arch/ppc/platforms/xilinx_ocp/xparameters_ml300.h and your system.mhs especially when looking at the interrupt signals and defines. If you made changes to xparameters_ml300.h you might want to rebuild the kernel completely: $ make clean dep bzImage modules.
Besides that I assume that you use a recent kernel (eg. the one in MVL3.1) and EDK (vs. V2PDK) as a development tool kit. - Peter Lou Rickard wrote: >And as the kernel boots, it loads an improved serial >module stored on the root filesystem, which works >fine? Would that explain why it works fine with >module support turned on, but doesn't work well with >module support turned off (and hence stuck using the >serial module that is built into the kernel)? > >~lr > >--- "VanBaren, Gerald (AGRE)" ><Gerald.VanBaren at smiths-aerospace.com> wrote: > > >>Just in case nobody answered yet (I didn't see an >>answer), the problem 99.9999% certainty is that your >>serial interrupts are not interrupting so the kernel >>is polling for serial I/O. This is a fallback >>operation with REALLY LONG timeouts. The symptoms >>are REALLY SLOW console interaction == your >>symptoms. Your follow-up message indicates the >>problem is in your serial driver module. >> >>gvb >> >> > > > > > > ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
