On 5/26/2022 2:55 AM, Pierre-Louis Bossart wrote:
+struct snd_soc_acpi_mach snd_soc_acpi_amd_rmb_sof_machines[] = {
+ {
+ .id = "AMDI1019",
+ .drv_name = "rmb-dsp",
+ .pdata = &acp_quirk_data,
+ .fw_filename = "sof-rmb.ri",
+ .sof_tplg_filename = "sof-acp-rmb.tplg",
+ },
+ {
+ .id = "10508825",
+ .drv_name = "nau8825-max",
+ .pdata = &acp_quirk_data,
+ .machine_quirk = snd_soc_acpi_codec_list,
+ .quirk_data = &_max,
+ .fw_filename = "sof-rmb.ri",
+ .sof_tplg_filename = "sof-acp-rmb.tplg",
this looks rather odd, you have two entries in the table that point to
the exact same pair of firmware and topology files. This is either
intentional and missing a comment, or a copy-paste mistake, or some of
these fields are not required.
This is intentional only ,will update this once platforms are confirmed,
will add comments
- clk_disable_unprepare(drvdata->wclk);
+ if (!drvdata->soc_mclk)
+ clk_disable_unprepare(drvdata->wclk);
}
mclk and wclk are different concepts usually.
Our intention here we don't want to enable /disable wclk(clock) when soc
is a clock master.
struct acp_card_drvdata {
@@ -49,6 +51,7 @@ struct acp_card_drvdata {
unsigned int dai_fmt;
struct clk *wclk;
struct clk *bclk;
+ bool soc_mclk;
I wonder if soc_mclk means 'soc_clock_provider' ?
yes
it looks like a configuration instead of a real/physical clock?
yes ,its just a flag to know whether soc is in a clock master mode or not.
