Can you try copy_to_user(), and not
__copy_to_user().

Kanoj

--- PRASENJIT CHAKRABORTY <pras_chakra@yahoo.com>

wrote:

quoted

Hello All,

   I posted this problem some days back but still
have
not received any solution.

So sorry for the resend again.

I am developing a driver. Though the driver is
functioning well but I am stuck up at the point of
transferring something from kernel to user space.

I am using __copy_to_user() and before calling

this

quoted

I
am checking the validity of user address with
verify_area().

Now the problem is that __copy_to_user() sometimes
return value > 0 which indicates a failure. This
happens everytime the user address is not

currently

quoted

present in the Physical Page i.e the when
__copy_to_user tries to copy to a page which is

not

quoted

currently paged in then it fails.

Moreover if I access the user buffer from my user
program before passing it to the kernel for

transfer

quoted

then __copy_to_user performs it happily or if I

lock

quoted

that page through mlock() system call. I read the
documentation but nowhere I found that I need to

do

quoted

something before __copy_to_user().

The problem for the time being has been

workarounded

quoted

by putting __verify_write() before __copy_to_user.

So I would like to know what is wrong with my
approach
or is it due to some other issue which is unknown

to

quoted

me (e.g corruption etc).

I shall be grateful to you all if you kindly help

me

quoted

out of this.

Thanks n Regards,

Prasenjit





__________________________________________________
Do You Yahoo!?
Make international calls for as low as $.04/minute
with Yahoo! Messenger
http://phonecard.yahoo.com/
--
To unsubscribe, send a message with 'unsubscribe
linux-mm' in
the body to majordomo@kvack.org.  For more info on
Linux MM,
see: http://www.linux-mm.org/

__________________________________________________
Do You Yahoo!?
Make international calls for as low as $.04/minute
with Yahoo! Messenger
http://phonecard.yahoo.com/
--
To unsubscribe, send a message with 'unsubscribe
linux-mm' in
the body to majordomo@kvack.org.  For more info on
Linux MM,
see: http://www.linux-mm.org/