Hello,

 

I am using a scatter gather DMA operation provided on the PPC44EPx to handle a large data move of 2MB from sixteen 128k kmalloc buffers.  I have sixteen scatter / gather descriptors pointing to the allocated memory which will be transferred to a peripheral on the EBC.  This all appears to be setup and working correctly.  Now I want to provide these 16 kmalloc buffers to a user space process and make them all appear as a contiguous 2MB buffer.

 

When the handler is called pageptr appears to be a valid page but the kernel traps with “Bad page state” “Trying to fix it up, but a reboot is needed” when returned from the handler.  Am I missing something required to modify the page state for the requested page?  Am I even handling the nopage request correctly?  I notice the nopage handler gets called for every 4096 byte page it is trying to map.  How do I map my 128k memory area into the 4k page requests?  Is there example code somewhere that demonstrates using a contiguous memory buffer in user space when it is comprised of multiple buffers in the kernel space?

 

 

First I am allocating 16 buffers using kmalloc:

 

#define DMABLOCKSIZE          1024 * 128

 

for (I = 0; I < 16; i++)

{

            device->dma_buf[i] = kmalloc(DMABLOCKSIZE, GFP_DMA | GFP_KERNEL);

            sgl->phyaddress = virt_to_phys(device->dma_buf[i]);

}

 

Then I am doing the following in the nopage handler:

 

page* nopage_handler(struct vma, ul address, int* type)

{

            unsigned long physaddr = address – vma->start;

            int index = physaddr / DMABLOCKSIZE;

            int off = physaddr % DMABLOCKSIZE;

            struct page* pageptr = virt_to_page(device->dma_buf[index] + off);

 

            get_page(pageptr);

            if (type)

                        *type = VM_FAULT_MINOR;

 

            return pageptr;

}

 

 

Dave Cogley

Software Engineer

Ultra Stereo Labs, Inc.

(805) 549-0161

mailto:dcogley@uslinc.com