Hi,
"Bartek Dolewski" <doles2@xxxxxxxxx> writes:
> Hi all
> Again I have problems with some kernel programming stuff. I decided to
> make my first device driver more robust. So I wanted to add
> access_ok() both in function read() and write() from fops. But when I
> type: echo "Message" >> /dev/mydevice I get this error:
> bash: echo: write error: Bad address
>
> Here my source code. First, implementation of read() function:
>
> static ssize_t FirstModule_Write(struct file *flip, const char
> *buffer, size_t length, loff_t *off)
> {
> int i;
>
> if(!access_ok(VERIFY_READ, buffer, length));
> return -EFAULT;
You return -EFAULT unconditionally ^.
No matter what access_ok() says.
It's funny that you already have it indented according to the actual
code-flow :)
> for(i=0; i < length && i < BUF_LEN; i++)
> get_user(msg2[i],buffer+i);
> msg2_Ptr = msg2;
> printk(KERN_INFO "Your message from Userland is: %s",msg2);
>
> I think that I have to verify reading mode because I`ll get some data
> from userspace buffer. Second argument is of course address of this
> buffer and last one ... that`s right, I don`t know how to interpret
> this argument. It is some length of memory block. There is the snippet
> from #include <asm/uaccess.h>:
> * access_ok: - Checks if a user space pointer is valid
> * @type: Type of access: %VERIFY_READ or %VERIFY_WRITE. Note that
> * %VERIFY_WRITE is a superset of %VERIFY_READ - if it is safe
> * to write to a block, it is always safe to read from it.
> * @addr: User space pointer to start of block to check
> * @size: Size of block to check
> Size of what block ? Size of address of this buffer or WHOLE buffer ?
The memory region that spans from addr to addr+size. So in your case
from buffer[0] to buffer[length - 1]. The call you have should be
correct if I didn't miss something.
Hannes
--
To unsubscribe from this list: send an email with
"unsubscribe kernelnewbies" to ecartis@xxxxxxxxxxxx
Please read the FAQ at http://kernelnewbies.org/FAQ
