| Message ID | 87va3y10a0.wl-kuninori.morimoto.gx@renesas.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [RFC] ASoC: simple-card-utils: fixup asoc_simple_card_get_dai_id() ID method | expand |
* Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> [181213 05:18]: > commit b6f3fc005a2c8 ("ASoC: simple-card-utils: fixup > asoc_simple_card_get_dai_id() counting") fixuped getting DAI ID method. > It will get DAI ID from OF graph "port", but, we want to consider about > "endpoint", too. > And, we also want to keep compatibility. > > This patch fixup it as > > if (driver has specified DAI ID) > use it as DAI ID > else if (OF graph endpoint has reg) > use it as DAI ID > else if (OF graph port has reg) > use it as DAI ID > else > use endpoint count as DAI ID Yup I tested this instead of the earlier 1/7 patch and things work nice for me. I also commented out reg properties from the codec end and the mapping worked fine. So please feel free to add: Tested-by: Tony Lindgren <tony@atomide.com>
Hi Tony Thank you for testing > > This patch fixup it as > > > > if (driver has specified DAI ID) > > use it as DAI ID > > else if (OF graph endpoint has reg) > > use it as DAI ID > > else if (OF graph port has reg) > > use it as DAI ID > > else > > use endpoint count as DAI ID > > Yup I tested this instead of the earlier 1/7 patch > and things work nice for me. I also commented out > reg properties from the codec end and the mapping > worked fine. So please feel free to add: > > Tested-by: Tony Lindgren <tony@atomide.com> Good !! Thank you, again !! I will re-post it as v3 patch-set Best regards --- Kuninori Morimoto
diff --git a/sound/soc/generic/simple-card-utils.c b/sound/soc/generic/simple-card-utils.c index 17d8aee..b807a47 100644 --- a/sound/soc/generic/simple-card-utils.c +++ b/sound/soc/generic/simple-card-utils.c @@ -269,22 +269,46 @@ EXPORT_SYMBOL_GPL(asoc_simple_card_parse_dai); static int asoc_simple_card_get_dai_id(struct device_node *ep) { + struct device_node *node; + struct device_node *endpoint; struct of_endpoint info; + int i, id; int ret; + /* use driver specified DAI ID if exist */ ret = snd_soc_get_dai_id(ep); if (ret != -ENOTSUPP) return ret; + /* use endpoint/port reg if exist */ + ret = of_graph_parse_endpoint(ep, &info); + if (ret == 0) { + if (info.id) + return info.id; + if (info.port) + return info.port; + } + + node = of_graph_get_port_parent(ep); + /* * Non HDMI sound case, counting port/endpoint on its DT * is enough. Let's count it. */ - ret = of_graph_parse_endpoint(ep, &info); - if (ret) - return -ENXIO; + i = 0; + id = -1; + for_each_endpoint_of_node(node, endpoint) { + if (endpoint == ep) + id = i; + i++; + } + + of_node_put(node); + + if (id < 0) + return -ENODEV; - return info.port; + return id; } int asoc_simple_card_parse_graph_dai(struct device_node *ep,
commit b6f3fc005a2c8 ("ASoC: simple-card-utils: fixup asoc_simple_card_get_dai_id() counting") fixuped getting DAI ID method. It will get DAI ID from OF graph "port", but, we want to consider about "endpoint", too. And, we also want to keep compatibility. This patch fixup it as if (driver has specified DAI ID) use it as DAI ID else if (OF graph endpoint has reg) use it as DAI ID else if (OF graph port has reg) use it as DAI ID else use endpoint count as DAI ID Fixes: commit b6f3fc005a2c8 ("ASoC: simple-card-utils: fixup asoc_simple_card_get_dai_id() counting") Reported-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> --- > Tony Can you test this patch ? It can keep compatibility, I think. > Mark I added [RFC] on this patch. If this patch was OK by Tony's test, I will re-post it as v3 patch-set. sound/soc/generic/simple-card-utils.c | 32 ++++++++++++++++++++++++++++---- 1 file changed, 28 insertions(+), 4 deletions(-)