On Wed, Oct 04, 2023 at 09:57:49AM -0400, Pierre-Louis Bossart wrote: > > > On 10/4/23 09:38, Greg Kroah-Hartman wrote: > > On Wed, Oct 04, 2023 at 09:16:47AM -0400, Pierre-Louis Bossart wrote: > >> > >> > >> On 10/4/23 09:02, Krzysztof Kozlowski wrote: > >>> If same devices with same device IDs are present on different soundwire > >>> buses, the probe fails due to conflicting device names and sysfs > >>> entries: > >>> > >>> sysfs: cannot create duplicate filename '/bus/soundwire/devices/sdw:0:0217:0204:00:0' > >>> > >>> The link ID is 0 for both devices, so they should be differentiated by > >>> bus ID. Add the bus ID so, the device names and sysfs entries look > >>> like: > >> > >> I am pretty sure this will break Intel platforms by changing the device > >> names. > >> > >> sof_sdw.c: else if (is_unique_device(adr_link, sdw_version, mfg_id, > >> part_id, > >> sof_sdw.c: > >> "sdw:%01x:%04x:%04x:%02x", link_id, > >> sof_sdw.c: > >> "sdw:%01x:%04x:%04x:%02x:%01x", link_id, > > > > device id name changes shouldn't break things, what is requring them to > > look a specific way? > > it's the ASoC dailink creation that relies on strings, we have similar > cases for I2C. > > There's no requirement that the name follows any specific convention, > just that when you want to rely on a specific device for an ASoC card > you need to use the string that matches its device name. matching the name is fine (if you are matching it against an existing name) but expecting the name to be anything specific is not going to work as the name is dynamic and can/will change each boot. thanks, greg k-h
