Re: [PATCH] remember to check return value from __copy_to_user() in cdrom_read_cdda_old()

[Jens Axboe <axboe@suse.de>]

On Tue, Sep 07 2004, viro@parcelfarce.linux.theplanet.co.uk wrote:
> On Tue, Sep 07, 2004 at 11:34:37AM +0200, Jens Axboe wrote:
> > On Tue, Sep 07 2004, Paul Mackerras wrote:
> > > Jens Axboe writes:
> > > 
> > > > __copy_to_user is the unchecking version of copy_to_user.
> > > 
> > > It doesn't range-check the address, but it does return non-zero
> > > (number of bytes not copied) if it encounters a fault writing to the
> > > user buffer.
> > 
> > but it doesn't matter, if it returns non-zero then something happened
> > between the access_ok() and the actual copy because the user app did
> > something silly. so I don't care much really, I think the major point is
> > the kernel will cope.
> > 
> > you could remove the access_ok() and change it to a copy_to_user()
> > instead, I don't care either way. it's the old and slow interface which
> > really never is used unless things have gone wrong anyways.
> 
> access_ok() is far from being the only reason for error here.  If area
> is unmapped, we shouldn't silently lose data without any indication of
> error.

it boils down to access_ok() not being sufficient on its own, and in
which case yes we should just use copy_to_user() and kill the check
completely as per the patch sent out.

-- 
Jens Axboe