Message ID | 20220818084216.1.I5c2b6fea19c4c0dec67fd4931f03df8e5ccaca95@changeid |
---|---|
State | New |
Headers | show |
Series | arm64: dts: qcom: sc7180: Add sleep state for alc5682 codec | expand |
Hi, On Thu, Aug 18, 2022 at 3:19 PM Joseph S. Barrera III <joebar@chromium.org> wrote: > > On 8/18/22 1:34 PM, Doug Anderson wrote: > > I tend to agree with Stephen's analysis. We actually need to keep the > > pullup enabled unless we are actually turning power off to the codec, > > which we don't seem to be doing. > > > > I guess I'm a little surprised that we don't even seem to turn any of > > this codec's regulators off in S3. That seems like it would be drawing > > power that we don't want. Maybe the "low power" mode of the codec is > > low enough and we need to avoid powering it off to avoid pops / hisses > > in S3 or something? If that's true, this might be one of those places > > where the "LPM" of the regulators might actually be useful... > > OK, fair enough, so suggestions on what I should do instead? Should I > look at why or how to turn the regulators off? Should I look into LPM? > Are there existing bugs for such work? It would be interesting to know from Judy if there's a reason we never turn the audio codec rails off. Maybe there's history that I'm not aware of? Matthias or Mengqi might also have ideas of how much power is at stake here? Ah, searching through the ChromeOS bug tracker finds some hits. At this point we should probably move the discussion off the lists and bring it back to the lists when we have some results. Unfortunately most of the bugs are not public and so having a discussion here is just noise for most people CCed. -Doug
On 8/18/22 3:39 PM, Doug Anderson wrote: > Ah, searching through the ChromeOS bug tracker finds some hits. At > this point we should probably move the discussion off the lists and > bring it back to the lists when we have some results. Unfortunately > most of the bugs are not public and so having a discussion here is > just noise for most people CCed. +1, thanks Doug
diff --git a/arch/arm64/boot/dts/qcom/sc7180-trogdor.dtsi b/arch/arm64/boot/dts/qcom/sc7180-trogdor.dtsi index b5f534db135a..94dd6c34d997 100644 --- a/arch/arm64/boot/dts/qcom/sc7180-trogdor.dtsi +++ b/arch/arm64/boot/dts/qcom/sc7180-trogdor.dtsi @@ -755,8 +755,9 @@ hp_i2c: &i2c9 { alc5682: codec@1a { compatible = "realtek,rt5682i"; reg = <0x1a>; - pinctrl-names = "default"; + pinctrl-names = "default", "sleep"; pinctrl-0 = <&hp_irq>; + pinctrl-1 = <&hp_sleep>; #sound-dai-cells = <1>; @@ -1336,6 +1337,18 @@ pinconf { }; }; + hp_sleep: hp-sleep { + pinmux { + pins = "gpio28"; + function = "gpio"; + }; + + pinconf { + pins = "gpio28"; + bias-disable; + }; + }; + pen_irq_l: pen-irq-l { pinmux { pins = "gpio21";
Add sleep state to acl5682. In default, gpio28 (HP_IRQ) is bias-pull-up. To save power, in the new sleep state, gpio28 is bias-disable. sleeping, /sys/kernel/debug/gpio shows gpio28 as "no pull". When codec is awake (microphone plugged in and in use), it shows gpio28 as "pull up". Signed-off-by: Joseph S. Barrera III <joebar@chromium.org> --- arch/arm64/boot/dts/qcom/sc7180-trogdor.dtsi | 15 ++++++++++++++- 1 file changed, 14 insertions(+), 1 deletion(-)