wm8962: error at soc_component_read_no_lock

wm8962: error at soc_component_read_no_lock

[Fabio Estevam]

Hi,

Running linux-next 20200714 on a imx6q sabresd I see the following errors:

wm8962 0-001a: ASoC: error at soc_component_read_no_lock on wm8962.0-001a: -16
wm8962 0-001a: ASoC: error at soc_component_read_no_lock on wm8962.0-001a: -16
 input: WM8962 Beep Generator as
/devices/platform/soc/2100000.bus/21a0000.i2c/i2c-0/0-001a/input/input3
wm8962 0-001a: ASoC: error at soc_component_read_no_lock on wm8962.0-001a: -16
wm8962 0-001a: ASoC: error at soc_component_read_no_lock on wm8962.0-001a: -16

Despite these errors, the codec probes fine.

What is the correct way to avoid such errors?

Thanks,

Fabio Estevam

Re: wm8962: error at soc_component_read_no_lock

[Kuninori Morimoto]

Hi Fabio

> wm8962 0-001a: ASoC: error at soc_component_read_no_lock on wm8962.0-001a: -16
> wm8962 0-001a: ASoC: error at soc_component_read_no_lock on wm8962.0-001a: -16
>  input: WM8962 Beep Generator as
> /devices/platform/soc/2100000.bus/21a0000.i2c/i2c-0/0-001a/input/input3
> wm8962 0-001a: ASoC: error at soc_component_read_no_lock on wm8962.0-001a: -16
> wm8962 0-001a: ASoC: error at soc_component_read_no_lock on wm8962.0-001a: -16
> 
> Despite these errors, the codec probes fine.
> 
> What is the correct way to avoid such errors?

New component function indicates above when error.
It seems yours couldn't read because of busy.
I think it had been happened, but didn't indicate it before.

Thank you for your help !!

Best regards
---
Kuninori Morimoto

Re: wm8962: error at soc_component_read_no_lock

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 969 bytes --]

On Fri, Jul 17, 2020 at 08:26:33AM +0900, Kuninori Morimoto wrote:
> > wm8962 0-001a: ASoC: error at soc_component_read_no_lock on wm8962.0-001a: -16
> > wm8962 0-001a: ASoC: error at soc_component_read_no_lock on wm8962.0-001a: -16
> >  input: WM8962 Beep Generator as
> > /devices/platform/soc/2100000.bus/21a0000.i2c/i2c-0/0-001a/input/input3
> > wm8962 0-001a: ASoC: error at soc_component_read_no_lock on wm8962.0-001a: -16
> > wm8962 0-001a: ASoC: error at soc_component_read_no_lock on wm8962.0-001a: -16

> > Despite these errors, the codec probes fine.

> > What is the correct way to avoid such errors?

> New component function indicates above when error.
> It seems yours couldn't read because of busy.
> I think it had been happened, but didn't indicate it before.

Right, IIRC the wm8962 had some registers that weren't accessible all
the time and possibly also some volatile bits - the driver will need
updates to handle things better for these widgets.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: wm8962: error at soc_component_read_no_lock

[Fabio Estevam]

On Fri, Jul 17, 2020 at 8:23 AM Mark Brown <broonie@kernel.org> wrote:

> Right, IIRC the wm8962 had some registers that weren't accessible all
> the time and possibly also some volatile bits - the driver will need
> updates to handle things better for these widgets.

Thanks, Mark.

I instrumented the code to print the register names so we can see
where soc_component_read_no_lock fails:

[    6.010526] ************ register is 512 (WM8962_GPIO_BASE)
[    6.014540] wm8962 0-001a: ASoC: error at
soc_component_read_no_lock on wm8962.0-001a: -16

[    6.030852] ************ register is 515 (This register does not exist)
[    6.034806] wm8962 0-001a: ASoC: error at
soc_component_read_no_lock on wm8962.0-001a: -16

[    6.162651] ************ register is 48 (WM8962_ADDITIONAL_CONTROL_4)
[    6.166524] wm8962 0-001a: ASoC: error at
soc_component_read_no_lock on wm8962.0-001a: -16
[    6.174911] ************ register is 48 (WM8962_ADDITIONAL_CONTROL_4)
[    6.178779] wm8962 0-001a: ASoC: error at
soc_component_read_no_lock on wm8962.0-001a: -16

Charles, Shengjiu, any suggestions, please?

Thanks

Re: wm8962: error at soc_component_read_no_lock

[Fabio Estevam]

On Fri, Jul 17, 2020 at 9:52 AM Fabio Estevam <festevam@gmail.com> wrote:
>
> On Fri, Jul 17, 2020 at 8:23 AM Mark Brown <broonie@kernel.org> wrote:
>
> > Right, IIRC the wm8962 had some registers that weren't accessible all
> > the time and possibly also some volatile bits - the driver will need
> > updates to handle things better for these widgets.
>
> Thanks, Mark.
>
> I instrumented the code to print the register names so we can see
> where soc_component_read_no_lock fails:
>
> [    6.010526] ************ register is 512 (WM8962_GPIO_BASE)
> [    6.014540] wm8962 0-001a: ASoC: error at
> soc_component_read_no_lock on wm8962.0-001a: -16
>
> [    6.030852] ************ register is 515 (This register does not exist)
> [    6.034806] wm8962 0-001a: ASoC: error at
> soc_component_read_no_lock on wm8962.0-001a: -16
>
> [    6.162651] ************ register is 48 (WM8962_ADDITIONAL_CONTROL_4)
> [    6.166524] wm8962 0-001a: ASoC: error at
> soc_component_read_no_lock on wm8962.0-001a: -16
> [    6.174911] ************ register is 48 (WM8962_ADDITIONAL_CONTROL_4)
> [    6.178779] wm8962 0-001a: ASoC: error at
> soc_component_read_no_lock on wm8962.0-001a: -16
>
> Charles, Shengjiu, any suggestions, please?

If I move WM8962_ADDITIONAL_CONTROL_4 to the non-volatile region, the
two errors related to this register are gone:

--- a/sound/soc/codecs/wm8962.c
+++ b/sound/soc/codecs/wm8962.c
@@ -149,6 +149,7 @@ static const struct reg_default wm8962_reg[] = {
        { 38, 0x0008 },   /* R38    - Right input PGA control */

        { 40, 0x0000 },   /* R40    - SPKOUTL volume */
+       { 48, 0x0000 },   /* R48    - Additional control(4) */
        { 41, 0x0000 },   /* R41    - SPKOUTR volume */

        { 49, 0x0010 },   /* R49    - Class D Control 1 */
@@ -790,7 +791,6 @@ static bool wm8962_volatile_register(struct device
*dev, unsigned int reg)
        case WM8962_CLOCKING1:
        case WM8962_SOFTWARE_RESET:
        case WM8962_THERMAL_SHUTDOWN_STATUS:
-       case WM8962_ADDITIONAL_CONTROL_4:
        case WM8962_DC_SERVO_6:
        case WM8962_INTERRUPT_STATUS_1:
        case WM8962_INTERRUPT_STATUS_2:

Is this a correct approach?

Thanks

Re: wm8962: error at soc_component_read_no_lock

[Fabio Estevam]

On Fri, Jul 17, 2020 at 10:01 AM Fabio Estevam <festevam@gmail.com> wrote:

> If I move WM8962_ADDITIONAL_CONTROL_4 to the non-volatile region, the
> two errors related to this register are gone:
>
> --- a/sound/soc/codecs/wm8962.c
> +++ b/sound/soc/codecs/wm8962.c
> @@ -149,6 +149,7 @@ static const struct reg_default wm8962_reg[] = {
>         { 38, 0x0008 },   /* R38    - Right input PGA control */
>
>         { 40, 0x0000 },   /* R40    - SPKOUTL volume */
> +       { 48, 0x0000 },   /* R48    - Additional control(4) */
>         { 41, 0x0000 },   /* R41    - SPKOUTR volume */
>
>         { 49, 0x0010 },   /* R49    - Class D Control 1 */
> @@ -790,7 +791,6 @@ static bool wm8962_volatile_register(struct device
> *dev, unsigned int reg)
>         case WM8962_CLOCKING1:
>         case WM8962_SOFTWARE_RESET:
>         case WM8962_THERMAL_SHUTDOWN_STATUS:
> -       case WM8962_ADDITIONAL_CONTROL_4:
>         case WM8962_DC_SERVO_6:
>         case WM8962_INTERRUPT_STATUS_1:
>         case WM8962_INTERRUPT_STATUS_2:
>
> Is this a correct approach?

I have sent a patch that fixes the soc_component_read_no_lock errors
for the GPIO registers.

I can send a formal patch for fixing soc_component_read_no_lock error
WM8962_ADDITIONAL_CONTROL_4 if this proposal looks good.

Thanks

Re: wm8962: error at soc_component_read_no_lock

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 431 bytes --]

On Fri, Jul 17, 2020 at 10:01:30AM -0300, Fabio Estevam wrote:

> If I move WM8962_ADDITIONAL_CONTROL_4 to the non-volatile region, the
> two errors related to this register are gone:

IIRC that register really did have some volatile bits in it, I can't
remember what they were or if they were important though.  It's not
fundamentally wrong and if the volatile bits don't practically matter
for Linux it would be sensible enough.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]