[Skiboot] [PATCH] pci: Add -Tn to SAS adapter's loc-code

[Gavin Shan]

On Fri, Mar 03, 2017 at 07:50:52AM +0530, Ananth N Mavinakayanahalli wrote:
>On Fri, Mar 03, 2017 at 10:54:42AM +1100, Gavin Shan wrote:
>> This adds -Tn to the device-tree node's loc-code property for SAS
>> adaptr as we did for ethernet adapter in commit 6558adff01bb ("
>> Add -Tn to ibm,loc-code for ethernet adaptors"). "n" represents
>> the function number of the PCI device.
>
>This would mean that the "n" would not necessarily be contiguous or
>start from 0 -- like what phyp does. Is that OK?
>

It's not changed by the patch. If it's not correct, what we expect
from here?

>> Signed-off-by: Gavin Shan <gwshan at linux.vnet.ibm.com>
>> ---
>>  core/pci.c | 4 +++-
>>  1 file changed, 3 insertions(+), 1 deletion(-)
>> 
>> diff --git a/core/pci.c b/core/pci.c
>> index 9889dbf..6c8f65c 100644
>> --- a/core/pci.c
>> +++ b/core/pci.c
>> @@ -1288,7 +1288,9 @@ static void pci_add_loc_code(struct dt_node *np, struct pci_device *pd)
>>  	/* XXX Don't do that on openpower for now, we will need to sort things
>>  	 * out later, otherwise the mezzanine slot on Habanero gets weird results
>>  	 */
>> -	if (class == 0x02 && sub == 0x00 && fsp_present()) {
>> +	if (fsp_present() &&
>> +	    ((class == 0x02 && sub == 0x00) ||	/* Ethernet */
>> +	     (class == 0x01 && sub == 0x07))) {	/* SAS      */
>>  		/* There's usually several spaces at the end of the property.
>>  		   Test for, but don't rely on, that being the case */
>>  		len = strlen(blcode);
>> -- 
>> 2.7.4
>> 
>> _______________________________________________
>> Skiboot mailing list
>> Skiboot at lists.ozlabs.org
>> https://lists.ozlabs.org/listinfo/skiboot