Re: [PATCH v2 02/10] cxl: Saperate out CXL dport->id vs actual dport hardware id

[Dave Jiang <dave.jiang@intel.com>]

On 5/12/25 10:04 PM, Gregory Price wrote:
> On Tue, May 06, 2025 at 05:43:02PM -0700, Dave Jiang wrote:
>> In preparation to allow dport to be allocated without being active, make
>> dport->id to be Linux id that enumerates the dport objects per port.
>> Keep the hardware id under dport->port_num to maintain compatibility and
>> introduce a dport->id as the enumeration id.
>>
> 
> This is more for the sake of clarity/documentation than anything else.
> 
> dport->port_num is the `hardware id` - i.e. what we'd find in an ACPI
> table or something.  This number may not necessarily be unique
> 
> dport->port_id is now a Linux ID that is unique to the device.

Yes. dport->id is guaranteed to be unique. While dport->port_num theoretically should be, but if a port is not active, it may have some bogus number in the register that clash with existing active port_num. I think Robert may have ran into this issue.

DJ


> 
> For example, on my existing test system, i have this
> 
> [/sys/bus/cxl/devices]# ls port1/dport*
> dport0  dport113  dport2 
> [/sys/bus/cxl/devices]# ls port2/dport*
> dport113  dport2
> [/sys/bus/cxl/devices]# ls port3/dport*
> dport0
> 
> I should now expect the following:
> 
> [/sys/bus/cxl/devices]# ls port1/dport*
> dport0  dport1  dport2 
> [/sys/bus/cxl/devices]# ls port2/dport*
> dport0  dport1
> [/sys/bus/cxl/devices]# ls port3/dport*
> dport0
> 
> 
> This should also apply to the root, whose dports were previously
> dictated by the host bridge numbers defined in the CHBS/DSDT.
> 
> [/sys/bus/cxl/devices/root0]# ls
> dport0  dport1  dport4  dport5
> 
> turns into
> 
> [/sys/bus/cxl/devices/root0]# ls
> dport0  dport1  dport2  dport3
> 
> 
> and my decoder0.0 target list ends up similarly changed
> 
> from
> [/sys/bus/cxl/devices/root0/decoder0.0]# cat target_list
> 5,4
> 
> to
> 
> [/sys/bus/cxl/devices/root0/decoder0.0]# cat target_list
> 3,2
> 
>> Signed-off-by: Dave Jiang <dave.jiang@intel.com>
> 
> If this is all correct, I like this.
> 
> Reviewed-by: Gregory Price <gourry@gourry.net>
> 
> ~Gregory