Re: [RFC PATCH 1/4] io_uring/splice: support do_splice_direct

[Christoph Hellwig]

On Thu, Nov 03, 2022 at 04:50:01PM +0800, Ming Lei wrote:
> do_splice_direct() has at least two advantages:
> 
> 1) the extra pipe isn't required from user viewpoint, so userspace
> code can be simplified, meantime easy to relax current pipe
> limit since curret->splice_pipe is used for direct splice
> 
> 2) in some situation, it isn't good to expose file data via
> ->splice_read() to userspace, such as the coming ublk driver's
> zero copy support, request pages will be spliced to pipe for
> supporting zero copy, and if it is READ, userspace may read
> data of kernel pages, and direct splice can avoid this kind
> of info leaks

Please make this a separate opcode instead of overloading the splice
op with a flag that causes very different behavior and isn't supported
for the regular splice syscall.