SuperI/O Access (Kernelspace)

[Jeremy Moles]

Hey guys, me again. :)

So, my device is working great now that I can access the GPIO pins via
userspace (using iotools w/ Tom Sylla's help) and can power on the
device. However, I'd like to be able to toggle the power on via a kernel
driver as well, but I'm having trouble accessing the same memory range
from kernelspace.

For example:

iotools io_read8 0xA00

Will return the state of the first 5 GPIO pins. I see that iotools calls
iopl() to be able to access this location from userspace.

When I try to do the same thing in the kernel:

unsigned short  b;
    unsigned short* ptr = (unsigned short*)(0xA00);

if(access_ok(VERIFY_READ, ptr, 8)) {
    	get_user(b, ptr);

...I immediately get a segfault. From what I've read and seen in other
example code, I believe I'm doing this right. However, I may be making
some wrong assumptions. I was just curious if anyone would be able to
shed some light on the subject--perhaps I'm missing some virtual memory
offset functions or similar...

[Tom Sylla]

Your address is in I/O address space, not memory address space. Take a
look at io_rw.c in iotools, you need to be using inb/outb (not just
reading/writing memory).

Tom

On Thu, Apr 14, 2011 at 4:12 PM, Jeremy Moles  wrote:
> Hey guys, me again. :)
>
> So, my device is working great now that I can access the GPIO pins via
> userspace (using iotools w/ Tom Sylla's help) and can power on the
> device. However, I'd like to be able to toggle the power on via a kernel
> driver as well, but I'm having trouble accessing the same memory range
> from kernelspace.
>
> For example:
>
> iotools io_read8 0xA00
>
> Will return the state of the first 5 GPIO pins. I see that iotools calls
> iopl() to be able to access this location from userspace.
>
> When I try to do the same thing in the kernel:
>
> unsigned short  b;
>        unsigned short* ptr = (unsigned short*)(0xA00);
>
> if(access_ok(VERIFY_READ, ptr, 8)) {
>                get_user(b, ptr);
>
> ...I immediately get a segfault. From what I've read and seen in other
> example code, I believe I'm doing this right. However, I may be making
> some wrong assumptions. I was just curious if anyone would be able to
> shed some light on the subject--perhaps I'm missing some virtual memory
> offset functions or similar...