2.4.9 odd problem

[Ken T. <ke...@we...>]

Hello,

A while ago I mentioned I was having an odd problem with 2.4.8.  A one
piece kernel with A4091 53c7xx.c driver was fine but a modular kernel with
the same driver hung up in the driver when booting even though the A4091
driver is not (and can't be) a module.

The problem went away with 2.4.9 but now has re-appeared.  I modified
virgefb.c CV64-3d driver, changed some of the ioremap calls and turned on
DEBUG, now both monolithic and modular kernels hang up in the A4091
initialization.

It looks like it hangs in ncr53c7xx_init() around

    page = __get_free_pages(GFP_ATOMIC,1);
    ...
    instance->hostdata[0] = page;
    memset((void *)instance->hostdata[0], 0, 8192);
    cache_push(virt_to_phys((void *)(instance->hostdata[0])), 8192);
    cache_clear(virt_to_phys((void *)(instance->hostdata[0])), 8192);
    kernel_set_cachemode(instance->hostdata[0], 8192, IOMAP_NOCACHE_SER);

I don't have much confidence poking around the cache code but will have a
go.

dmesg attached.

Ken.