Re: virtio(-scsi) vs. chained sg_lists (was Re: [PATCH] scsi: virtio-scsi: Fix address translation failure of HighMem pages used by sg list)

[Paolo Bonzini <pbonzini@redhat.com>]

Il 25/07/2012 21:16, Boaz Harrosh ha scritto:
> The picture confused me. It looked like the first element is the virtio_scsi_cmd_req
> not an sgilist-element that points to the struct's buffer.
> 
> In that case then yes your plan of making a two-elements fragment that points to the
> original scsi-sglist is perfect. All you have to do is that, and all you have to do
> at virtio is use the sg_for_each macro and you are done.
> 
> You don't need any sglist allocation or reshaping. And you can easily support
> chaining. Looks like order of magnitude more simple then what you do now

It is.

> So what is the problem?

That not all architectures have ARCH_HAS_SG_CHAIN (though all those I
care about do).  So I need to go through all architectures and make sure
they use for_each_sg, or at least to change ARCH_HAS_SG_CHAIN to a
Kconfig define so that dependencies can be expressed properly.

> And BTW you won't need that new __sg_set_page API anymore.

Kind of.

   sg_init_table(sg, 2);
   sg_set_buf(sg[0], req, sizeof(req));
   sg_chain(sg[1], scsi_out(sc));

is still a little bit worse than

   __sg_set_buf(sg[0], req, sizeof(req));
   __sg_chain(sg[1], scsi_out(sc));

Paolo