Re: [PATCH v5 07/10] PCI/IDE: Add IDE establishment helpers

[<dan.j.williams@intel.com>]

Alexey Kardashevskiy wrote:
[..]
> > diff --git a/include/linux/pci-ide.h b/include/linux/pci-ide.h
> > index cf1f7a10e8e0..a2d3fb4a289b 100644
> > --- a/include/linux/pci-ide.h
> > +++ b/include/linux/pci-ide.h
> > @@ -24,6 +24,8 @@ enum pci_ide_partner_select {
> >    * @rid_start: Partner Port Requester ID range start
> >    * @rid_start: Partner Port Requester ID range end
> >    * @stream_index: Selective IDE Stream Register Block selection
> > + * @default_stream: Endpoint uses this stream for all upstream TLPs regardless of
> > + * 		    address and RID association registers
> >    * @setup: flag to track whether to run pci_ide_stream_teardown() for this
> >    *	   partner slot
> >    * @enable: flag whether to run pci_ide_stream_disable() for this partner slot
> > @@ -32,6 +34,7 @@ struct pci_ide_partner {
> >   	u16 rid_start;
> >   	u16 rid_end;
> >   	u8 stream_index;
> > +	unsigned int default_stream:1;
> 
> 
> This sets "Default" on both ends and the rootport does not need it in
> my setup (it does not seem to affect anything though) - the  rootport
> always knows the stream ID from the RMP entry of a MMIO being
> accessed. May be move it to pci_ide_partner? Thanks,

I was about to change this, but it is already the case that "default
stream" is only an Endpoint port concept in pairing of endpoint and root
port. Also, this is already in 'struct pci_ide_partner'. So if your TSM
driver sets this for the PCI_IDE_RP case, then that is your TSM driver
concern, I do not think the core should worry about filtering mistaken
settings from TSM drivers.