Mapping physical addresses to virtual address linux

Question

I am working on a small embedded system. When my linux boots up into user space, I know where are my devices in the physical memory. I want to map them into user space virtual addresses. Currently, I am doing it through a kernel module. I use vmalloc/kmalloc (depending on the size) and then I use ioremap_page_range on that returned virtual addresses to map my physical addresses. I dont think that is the correct way to go about. First of all I am allocating memory and then I am asking kernel to remap that virtual address space to some different physical address space. (Initially mapped physical->virtual in vmcall/kmalloc is kinda useless as I dont care about those physical pages. This is definitely not good.)

Instead of this is there a better way to map the known physical memory into user space process. (I know other than my user space process, no one gonna touch that memory.)

Thanks

Answer 1

What you are trying to do is accessing what is called IO memory. I can only encourage you to read the Linux Device Drivers (LDD) book and more specifically the chapter 9.

To "allocate" such an area, you need to call

struct resource *request_mem_region(unsigned long start, unsigned long len, char *name)

. Before your driver can access it, you have to assign it a virtual address, this is done with a call to

void *ioremap(unsigned long phys_addr, unsigned long size)

To ensure that your driver will then work on different architectures/platforms, be sure to use some accessor function to such areas ( ioread8/16/32 or iowrite8/16/32 and all of their variants).

Answer 2

In Kernel module, remap_pfn_range() can be used to convert the physical address to virtual address. The following link will be helpful.

How remap_pfn_range remaps kernel memory to user space?

Answer 3

In Kernel module, remap_pfn_range() can be used to convert the physical address to virtual address. When you don't have a actual devices you can: 1) create a virtual device and, 2) use mmap to those virtual devices to access the very same kernel memory through remap_pfn_range virtual mapping of that process. 3) Usually in dedicated environments you may addition want to pin those physical pages lest they are taken away from your process. 4) You also share these physical addresses with different processes but will need to handle synchronization, independently through other IPC mechanisms as to each process they will look as different addresses.