On Thu, Jul 27, 2023 at 03:34:53PM +0800, Tim Jiang wrote: > On 7/27/23 15:27, Johan Hovold wrote: > > On Thu, Jul 27, 2023 at 12:40:10PM +0800, Tim Jiang wrote: > >> @@ -1762,10 +1763,32 @@ static int qca_setup(struct hci_uart *hu) > >> */ > >> set_bit(HCI_QUIRK_SIMULTANEOUS_DISCOVERY, &hdev->quirks); > >> > >> - bt_dev_info(hdev, "setting up %s", > >> - qca_is_wcn399x(soc_type) ? "wcn399x" : > >> - (soc_type == QCA_WCN6750) ? "wcn6750" : > >> - (soc_type == QCA_WCN6855) ? "wcn6855" : "ROME/QCA6390"); > >> + switch (soc_type) { > >> + case QCA_AR3002: > >> + soc_name = "ar300x"; > >> + break; > >> + case QCA_ROME: > >> + soc_name = "ROME"; > >> + break; > >> + case QCA_QCA6390: > >> + soc_name = "QCA6390"; > >> + break; > >> + case QCA_WCN3990: > >> + case QCA_WCN3991: > >> + case QCA_WCN3998: > >> + soc_name = "wcn399x"; > >> + break; > >> + case QCA_WCN6750: > >> + soc_name = "wcn6750"; > >> + break; > >> + case QCA_WCN6855: > >> + soc_name = "wcn6855"; > >> + break; > > I still think the above should be sorted (alphabetically) as maintaining > > these lists otherwise soon becomes harder than it should be. And similar > > throughout the driver. > [Tim] Hi Johan: I think we no need to sort it, we only add the new btsoc > name following the older one, for example , ar300x is the oldest , ROME > is new than ar300x, actually qca2066 is newer version chip than qca6390, > so I does not think we need to sort it. Possibly, but generally this becomes hard to maintain and eventually someone will need to sort these entries anyway. Therefore it's generally a good idea to just do so from the start. But it was good that you replied so that we know that this comment was not just missed or ignored. > >> + default: > >> + soc_name = "unknown soc"; > >> + break; > >> + } > >> + bt_dev_info(hdev, "setting up %s", soc_name); Johan