Re: aio_read const struct iovec *

[Matthew Wilcox]

On Wed, Sep 09, 2009 at 02:18:38PM +0300, Sergey Senozhatsky wrote:
> Hello,
> I have the following question:
> struct file_operations {
> 	ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
> 	ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
> At the same time:
> 	ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
> 	ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
> 
> 'const struct iovec *' both for aio_write and aio_read. 
> And "char __user *" and "const char __user *" for read/write.
> 
> For example,
> pipe_read(struct kiocb *iocb, const struct iovec *_iov,
> 	    unsigned long nr_segs, loff_t ppos)
> {
> 	struct iovec *iov = (struct iovec *)_iov; 
> 
> Could we avoid 'struct iovec *iov = (struct iovec *)_iov;' by changing aio_read 'const struct iovec *'
>  to 'struct iovec *'?

If you try it, it'll be educational for you to see where it fails ;-)

But if you're too impatient, aio_rw_vect_retry() will be (one of)
the problems:

        ssize_t (*rw_op)(struct kiocb *, const struct iovec *,
                         unsigned long, loff_t);
[...]
                rw_op = file->f_op->aio_read;
[...]
                rw_op = file->f_op->aio_write;


so aio_read and aio_write need to have the same prototype.

Now, I'm not sure why it's more beneficial to have a lying const in the
aio_read prototype than simply omitting the const from the aio_write
prototype ... but I'm sure if you try that, you'll find a good reason too.

-- 
Matthew Wilcox				Intel Open Source Technology Centre
"Bill, look, we understand that you're interested in selling us this
operating system, but compare it to ours.  We can't possibly take such
a retrograde step."
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/