Re: [PATCH v3 net-next 01/21] iov_iter: Introduce new procedures for copy to iter/pages

[Or Gerlitz <gerlitz.or@gmail.com>]

On Mon, Feb 1, 2021 at 7:38 PM Christoph Hellwig <hch@lst.de> wrote:
> On Mon, Feb 01, 2021 at 12:04:49PM +0200, Boris Pismenny wrote:
> > +static __always_inline __must_check
> > +size_t ddp_copy_to_iter(const void *addr, size_t bytes, struct iov_iter *i)
> > +{
> > +     if (unlikely(!check_copy_size(addr, bytes, true)))
> > +             return 0;
> > +     else
> > +             return _ddp_copy_to_iter(addr, bytes, i);
> > +}
>
> No need for the else after a return, and the normal kernel convention
> double underscores for magic internal functions.

ack for the no-else-after-a-return

Re the double underscoring, I was not sure, e.g the non-ddp counterpart
(_copy_to_iter) is single underscored

> But more importantly: does this belong into the generic header without
> and comments what the ddp means and when it should be used?

will look into this, any idea for a more suitable location?

> > +static void ddp_memcpy_to_page(struct page *page, size_t offset, const char *from, size_t len)
>
> Overly long line.  But we're also looking into generic helpers for
> this kind of things, not sure if they made it to linux-next in the
> meantime, but please check.

This is what I found in linux-next - note sure if you were referring to it

commit 11432a3cc061c39475295be533c3674c4f8a6d0b
Author: David Howells <dhowells@redhat.com>

    iov_iter: Add ITER_XARRAY

> > +size_t _ddp_copy_to_iter(const void *addr, size_t bytes, struct iov_iter *i)
> > +{
> > +     const char *from = addr;
> > +     if (unlikely(iov_iter_is_pipe(i)))
> > +             return copy_pipe_to_iter(addr, bytes, i);
> > +     if (iter_is_iovec(i))
> > +             might_fault();
> > +     iterate_and_advance(i, bytes, v,
> > +             copyout(v.iov_base, (from += v.iov_len) - v.iov_len, v.iov_len),
> > +             ddp_memcpy_to_page(v.bv_page, v.bv_offset,
> > +                                (from += v.bv_len) - v.bv_len, v.bv_len),
> > +             memcpy(v.iov_base, (from += v.iov_len) - v.iov_len, v.iov_len)
> > +             )
> > +
> > +     return bytes;
> > +}
>
> This bloats every kernel build, so please move it into a conditionally built file.

ack

>  And please document the whole thing.

ok