Re: [PATCH v6 1/3] ublk: add opcode offsets for DRV_IN/DRV_OUT

[Ming Lei <ming.lei@redhat.com>]

On Sun, Jul 09, 2023 at 11:52:39PM -0700, Christoph Hellwig wrote:
> On Fri, Jul 07, 2023 at 08:59:03AM +0800, Ming Lei wrote:
> > > let's clearly state so. But then, I still not understand why these need
> > > a different naming pattern using the "__UBLK" prefix...
> > 
> > I think __UBLK just meant we don't suggest userspace to use it directly,
> > since the added macros are just for making ranges for DRV_IN and DRV_OUT,
> > so we can check command direction easily be using this start/end info in
> > both sides.
> 
> Folks, please stop coupling a uapi (or on-disk protocol) too tightly
> to Linux internals.  Think of what makes sense as a communication
> protocol, not what is an internal kernel interface.
> 
> REPORT_ZONES is a sensible command, and supported in ATA/SCSI/NVMe in
> one way or another.  In Linux it is a synchronous method call right now
> for one reason or another, and most implementation map it to a
> passthrough command - be that the actual protocol command or something
> internal for virtio.
> 
> So for ublk this is just another command like any other, that needs to
> be defined and documented.  Nothing internal or driver specific.
 
Yes, that is exactly what we are doing.

The added macros of UBLK_IO_OP_DRV_IN_START[END] are just for supporting
more ublk passthrough commands, and the motivation is for running
check(such as buffer direction) in two sides easily.

However, I think it is just fine to delay to add it until introducing
the 2nd ublk pt command.

Thanks, 
Ming