[BUG] On Allwinner A20 Board Cubietruck distorted SPDIF opical output

[BUG] On Allwinner A20 Board Cubietruck distorted SPDIF opical output

['Markus Mitsch' via linux-sunxi]

Hello,

i have a new problem. if i want to output sound from optical i get
distorted sound. At 44.1 khz its like an old vinly, but at 48 khz and
higher it is very bad. thanks in advance for your time.

mfg

Re: [BUG] On Allwinner A20 Board Cubietruck distorted SPDIF opical output

[Code Kipper]

On 20 May 2018 at 14:23, 'Markus Mitsch' via linux-sunxi
<linux-sunxi-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org> wrote:
> Hello,
>
> i have a new problem. if i want to output sound from optical i get
> distorted sound. At 44.1 khz its like an old vinly, but at 48 khz and
> higher it is very bad. thanks in advance for your time.
Hi Markus,
I haven't tested it for a while but when I did it worked as high as
96K with no issues. Does it work ok with the original cubietruck
software?,
CK
>
> mfg
>
> --
> You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
> For more options, visit https://groups.google.com/d/optout.

Re: [BUG] On Allwinner A20 Board Cubietruck distorted SPDIF opical output

['Markus Mitsch' via linux-sunxi]

>> Hello,
>>
>> i have a new problem. if i want to output sound from optical i get
>> distorted sound. At 44.1 khz its like an old vinly, but at 48 khz and
>> higher it is very bad. thanks in advance for your time.

> Hi Markus,
> I haven't tested it for a while but when I did it worked as high as
> 96K with no issues. Does it work ok with the original cubietruck
> software?,
> CK

Hello,
i am using arch linux with mainline kernel 4.16.8. parhaps i can try
today with official image.
mfg

Re: [BUG] On Allwinner A20 Board Cubietruck distorted SPDIF opical output

['Markus Mitsch' via linux-sunxi]

2018-05-21 11:43 GMT+02:00 Markus Mitsch <markusmitsch86-gM/Ye1E23mwN+BqQ9rBEUg@public.gmane.org>:
>>> Hello,
>>>
>>> i have a new problem. if i want to output sound from optical i get
>>> distorted sound. At 44.1 khz its like an old vinly, but at 48 khz and
>>> higher it is very bad. thanks in advance for your time.
>
>> Hi Markus,
>> I haven't tested it for a while but when I did it worked as high as
>> 96K with no issues. Does it work ok with the original cubietruck
>> software?,
>> CK
>
> Hello,
> i am using arch linux with mainline kernel 4.16.8. parhaps i can try
> today with official image.
> mfg

Hi.
So, with debian image from 2014 with 3.4.79 kernel the sound is clear
up to 96 khz.
mfg

Re: [BUG] On Allwinner A20 Board Cubietruck distorted SPDIF opical output

[Maxime Ripard]

[-- Attachment #1: Type: text/plain, Size: 1012 bytes --]

On Mon, May 21, 2018 at 02:43:55PM +0200, Markus Mitsch wrote:
> 2018-05-21 11:43 GMT+02:00 Markus Mitsch <markusmitsch86-gM/Ye1E23mwN+BqQ9rBEUg@public.gmane.org>:
> >>> Hello,
> >>>
> >>> i have a new problem. if i want to output sound from optical i get
> >>> distorted sound. At 44.1 khz its like an old vinly, but at 48 khz and
> >>> higher it is very bad. thanks in advance for your time.
> >
> >> Hi Markus,
> >> I haven't tested it for a while but when I did it worked as high as
> >> 96K with no issues. Does it work ok with the original cubietruck
> >> software?,
> >> CK
> >
> > Hello,
> > i am using arch linux with mainline kernel 4.16.8. parhaps i can try
> > today with official image.
> > mfg
> 
> Hi.
> So, with debian image from 2014 with 3.4.79 kernel the sound is clear
> up to 96 khz.

You said that it is a new problem, when was it introduced? Could you
bisect the issue?

Maxime

-- 
Maxime Ripard, Bootlin (formerly Free Electrons)
Embedded Linux and Kernel engineering
https://bootlin.com

Re: [BUG] On Allwinner A20 Board Cubietruck distorted SPDIF opical output

['Markus Mitsch' via linux-sunxi]

2018-05-21 15:17 GMT+02:00 Maxime Ripard <maxime.ripard-LDxbnhwyfcJBDgjK7y7TUQ@public.gmane.org>:
> On Mon, May 21, 2018 at 02:43:55PM +0200, Markus Mitsch wrote:
>> 2018-05-21 11:43 GMT+02:00 Markus Mitsch <markusmitsch86-gM/Ye1E23mwN+BqQ9rBEUg@public.gmane.org>:
>> >>> Hello,
>> >>>
>> >>> i have a new problem. if i want to output sound from optical i get
>> >>> distorted sound. At 44.1 khz its like an old vinly, but at 48 khz and
>> >>> higher it is very bad. thanks in advance for your time.
>> >
>> >> Hi Markus,
>> >> I haven't tested it for a while but when I did it worked as high as
>> >> 96K with no issues. Does it work ok with the original cubietruck
>> >> software?,
>> >> CK
>> >
>> > Hello,
>> > i am using arch linux with mainline kernel 4.16.8. parhaps i can try
>> > today with official image.
>> > mfg
>>
>> Hi.
>> So, with debian image from 2014 with 3.4.79 kernel the sound is clear
>> up to 96 khz.
>
> You said that it is a new problem, when was it introduced? Could you
> bisect the issue?
>
> Maxime

Hello.
I dont know since when this problem appears. When saying "new" i
refered to my older post. I tried to look into the dts file and the
code
but i am afraid this is a bit too high for me. I can say that the
analog output works, even at high sampling rates.
And I have read something about a clock source could be the problem.
I'm sorry I can't tell you any more.

mfg
Markus

Re: [BUG] On Allwinner A20 Board Cubietruck distorted SPDIF opical output

[Code Kipper]

Hi Markus,
I've been able to reproduce this on my A31 board with linux-next(plays
at 44100 but not at 96K)so it is an issue. I also tested the same code
on my A64 and it worked fine, The blocks are very similar so it maybe
a clocking issue, I'll look into this.
BR,
CK

On 21 May 2018 at 20:17, Markus Mitsch <markusmitsch86-gM/Ye1E23mwN+BqQ9rBEUg@public.gmane.org> wrote:
> 2018-05-21 15:17 GMT+02:00 Maxime Ripard <maxime.ripard-LDxbnhwyfcJBDgjK7y7TUQ@public.gmane.org>:
>> On Mon, May 21, 2018 at 02:43:55PM +0200, Markus Mitsch wrote:
>>> 2018-05-21 11:43 GMT+02:00 Markus Mitsch <markusmitsch86-gM/Ye1E23mwN+BqQ9rBEUg@public.gmane.org>:
>>> >>> Hello,
>>> >>>
>>> >>> i have a new problem. if i want to output sound from optical i get
>>> >>> distorted sound. At 44.1 khz its like an old vinly, but at 48 khz and
>>> >>> higher it is very bad. thanks in advance for your time.
>>> >
>>> >> Hi Markus,
>>> >> I haven't tested it for a while but when I did it worked as high as
>>> >> 96K with no issues. Does it work ok with the original cubietruck
>>> >> software?,
>>> >> CK
>>> >
>>> > Hello,
>>> > i am using arch linux with mainline kernel 4.16.8. parhaps i can try
>>> > today with official image.
>>> > mfg
>>>
>>> Hi.
>>> So, with debian image from 2014 with 3.4.79 kernel the sound is clear
>>> up to 96 khz.
>>
>> You said that it is a new problem, when was it introduced? Could you
>> bisect the issue?
>>
>> Maxime
>
> Hello.
> I dont know since when this problem appears. When saying "new" i
> refered to my older post. I tried to look into the dts file and the
> code
> but i am afraid this is a bit too high for me. I can say that the
> analog output works, even at high sampling rates.
> And I have read something about a clock source could be the problem.
> I'm sorry I can't tell you any more.
>
> mfg
> Markus

Re: [BUG] On Allwinner A20 Board Cubietruck distorted SPDIF opical output

[Code Kipper]

Hi All,
I can see that the problem was introduced somewhere here

DOESN'T WORK
ffc3eb6f3a83 Merge tag 'sunxi-clk-for-4.15' of
https://git.kernel.org/pub/scm/linux/kernel/git/sunxi/linux into
clk-next
ae74ac082886 Merge tag 'clk-v4.15-exynos-pm' of
git://git.kernel.org/pub/scm/linux/kernel/git/snawrocki/clk into
clk-next
faa865f18cb7 Merge tag 'clk-v4.15-samsung' of
git://git.kernel.org/pub/scm/linux/kernel/git/snawrocki/clk into
clk-next
f09a6b86fdae Merge tag 'clk-renesas-for-v4.15-tag1' of
git://git.kernel.org/pub/scm/linux/kernel/git/geert/renesas-drivers
into clk-next
4328a2186e51 clk: sunxi-ng: sun4i: Export video PLLs
553c7d5ba2fe clk: sunxi-ng: Add A83T display clocks
7679eb20353d clk: samsung: Add a separate driver for Exynos4412 ISP clocks
8ca8ac102484 clk: samsung: Add dt bindings for Exynos4412 ISP clock controller
75920aac275a clk: samsung: Instantiate Exynos4412 ISP clocks only when available
24ea78a09f89 clk: sunxi-ng: sun8i: a23: Use sigma-delta modulation for audio PLL
ee6501d69217 clk: sunxi-ng: sun6i: Use sigma-delta modulation for audio PLL
042f7f8f9715 clk: sunxi-ng: sun5i: Use sigma-delta modulation for audio PLL
de3448519194 clk: sunxi-ng: sun4i: Use sigma-delta modulation for audio PLL
a5e3e2b2ef85 clk: sunxi-ng: sun8i: h3: Use sigma-delta modulation for audio PLL
392ba5fafcdf clk: sunxi-ng: nm: Add support for sigma-delta modulation
05d2eaac96d4 clk: sunxi-ng: Add sigma-delta modulation support
4cdbc40d64d4 clk: sunxi-ng: nm: Check if requested rate is supported
by fractional clock
d51fe3ba9773 clk: sunxi-ng: sun5i: Fix bit offset of audio PLL post-divider
4e8975cbb516 clk: samsung: exynos5433: mark PM functions as __maybe_unused
a4f21e9ceb5c clk: samsung: Remove obsolete clkdev alias support
45d882daf88a clk: samsung: Add explicit MPLL, EPLL clkdev aliases in
S3C2443 driver
cd05417f728b clk: samsung: Rework clkdev alias handling in S3C2443 driver
efea8d377184 clk: samsung: Rework clkdev alias handling in Exynos5440 driver
36ba48240b19 clk: samsung: Drop useless alias in Exynos5420 clk driver
29964890f31c clk: samsung: Remove clkdev alias support in Exynos5250 clk driver
6de08891c896 clk: samsung: Remove double assignment of CLK_ARM_CLK in
Exynos4 driver
58f4a5ff3a62 clk: samsung: Remove clkdev alias support in Exynos4 clk driver
c9194fb623b0 clk: samsung: Remove support for obsolete Exynos4212 CPU clock
d5cd103b06f9 clk: samsung: Remove support for Exynos4212 SoCs in
Exynos CLKOUT driver
7c4f63ec94a1 clk: sunxi-ng: a83t: Fix invalid csi-mclk mux offset
WORKS
0a4e632b6f9d Merge branch 'clk-fixes' into clk-next

BR,
CK


On 22 May 2018 at 21:27, Code Kipper <codekipper-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> Hi Markus,
> I've been able to reproduce this on my A31 board with linux-next(plays
> at 44100 but not at 96K)so it is an issue. I also tested the same code
> on my A64 and it worked fine, The blocks are very similar so it maybe
> a clocking issue, I'll look into this.
> BR,
> CK
>
> On 21 May 2018 at 20:17, Markus Mitsch <markusmitsch86-gM/Ye1E23mwN+BqQ9rBEUg@public.gmane.org> wrote:
>> 2018-05-21 15:17 GMT+02:00 Maxime Ripard <maxime.ripard-LDxbnhwyfcJBDgjK7y7TUQ@public.gmane.org>:
>>> On Mon, May 21, 2018 at 02:43:55PM +0200, Markus Mitsch wrote:
>>>> 2018-05-21 11:43 GMT+02:00 Markus Mitsch <markusmitsch86-gM/Ye1E23mwN+BqQ9rBEUg@public.gmane.org>:
>>>> >>> Hello,
>>>> >>>
>>>> >>> i have a new problem. if i want to output sound from optical i get
>>>> >>> distorted sound. At 44.1 khz its like an old vinly, but at 48 khz and
>>>> >>> higher it is very bad. thanks in advance for your time.
>>>> >
>>>> >> Hi Markus,
>>>> >> I haven't tested it for a while but when I did it worked as high as
>>>> >> 96K with no issues. Does it work ok with the original cubietruck
>>>> >> software?,
>>>> >> CK
>>>> >
>>>> > Hello,
>>>> > i am using arch linux with mainline kernel 4.16.8. parhaps i can try
>>>> > today with official image.
>>>> > mfg
>>>>
>>>> Hi.
>>>> So, with debian image from 2014 with 3.4.79 kernel the sound is clear
>>>> up to 96 khz.
>>>
>>> You said that it is a new problem, when was it introduced? Could you
>>> bisect the issue?
>>>
>>> Maxime
>>
>> Hello.
>> I dont know since when this problem appears. When saying "new" i
>> refered to my older post. I tried to look into the dts file and the
>> code
>> but i am afraid this is a bit too high for me. I can say that the
>> analog output works, even at high sampling rates.
>> And I have read something about a clock source could be the problem.
>> I'm sorry I can't tell you any more.
>>
>> mfg
>> Markus

Re: [BUG] On Allwinner A20 Board Cubietruck distorted SPDIF opical output

['Markus Mitsch' via linux-sunxi]

hi.

i posted my answer wrong so here is my finding:
i made the output work by disabling the sigma delta modulation.
after compiling the kernel  i can output through spdif with any rate.

greetings
markus

2018-05-26 18:38 GMT+02:00 Markus Mitsch <markusmitsch86-gM/Ye1E23mwN+BqQ9rBEUg@public.gmane.org>:
> 2018-05-26 16:57 GMT+02:00 Markus Mitsch <markusmitsch86-gM/Ye1E23mwN+BqQ9rBEUg@public.gmane.org>:
>> 2018-05-23 17:12 GMT+02:00 Code Kipper <codekipper-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>:
>>> Hi All,
>>> I can see that the problem was introduced somewhere here
>>>
>>> DOESN'T WORK
>>> ffc3eb6f3a83 Merge tag 'sunxi-clk-for-4.15' of
>>> https://git.kernel.org/pub/scm/linux/kernel/git/sunxi/linux into
>>> clk-next
>>> ae74ac082886 Merge tag 'clk-v4.15-exynos-pm' of
>>> git://git.kernel.org/pub/scm/linux/kernel/git/snawrocki/clk into
>>> clk-next
>>> faa865f18cb7 Merge tag 'clk-v4.15-samsung' of
>>> git://git.kernel.org/pub/scm/linux/kernel/git/snawrocki/clk into
>>> clk-next
>>> f09a6b86fdae Merge tag 'clk-renesas-for-v4.15-tag1' of
>>> git://git.kernel.org/pub/scm/linux/kernel/git/geert/renesas-drivers
>>> into clk-next
>>> 4328a2186e51 clk: sunxi-ng: sun4i: Export video PLLs
>>> 553c7d5ba2fe clk: sunxi-ng: Add A83T display clocks
>>> 7679eb20353d clk: samsung: Add a separate driver for Exynos4412 ISP clocks
>>> 8ca8ac102484 clk: samsung: Add dt bindings for Exynos4412 ISP clock controller
>>> 75920aac275a clk: samsung: Instantiate Exynos4412 ISP clocks only when available
>>> 24ea78a09f89 clk: sunxi-ng: sun8i: a23: Use sigma-delta modulation for audio PLL
>>> ee6501d69217 clk: sunxi-ng: sun6i: Use sigma-delta modulation for audio PLL
>>> 042f7f8f9715 clk: sunxi-ng: sun5i: Use sigma-delta modulation for audio PLL
>>> de3448519194 clk: sunxi-ng: sun4i: Use sigma-delta modulation for audio PLL
>>> a5e3e2b2ef85 clk: sunxi-ng: sun8i: h3: Use sigma-delta modulation for audio PLL
>>> 392ba5fafcdf clk: sunxi-ng: nm: Add support for sigma-delta modulation
>>> 05d2eaac96d4 clk: sunxi-ng: Add sigma-delta modulation support
>>> 4cdbc40d64d4 clk: sunxi-ng: nm: Check if requested rate is supported
>>> by fractional clock
>>> d51fe3ba9773 clk: sunxi-ng: sun5i: Fix bit offset of audio PLL post-divider
>>> 4e8975cbb516 clk: samsung: exynos5433: mark PM functions as __maybe_unused
>>> a4f21e9ceb5c clk: samsung: Remove obsolete clkdev alias support
>>> 45d882daf88a clk: samsung: Add explicit MPLL, EPLL clkdev aliases in
>>> S3C2443 driver
>>> cd05417f728b clk: samsung: Rework clkdev alias handling in S3C2443 driver
>>> efea8d377184 clk: samsung: Rework clkdev alias handling in Exynos5440 driver
>>> 36ba48240b19 clk: samsung: Drop useless alias in Exynos5420 clk driver
>>> 29964890f31c clk: samsung: Remove clkdev alias support in Exynos5250 clk driver
>>> 6de08891c896 clk: samsung: Remove double assignment of CLK_ARM_CLK in
>>> Exynos4 driver
>>> 58f4a5ff3a62 clk: samsung: Remove clkdev alias support in Exynos4 clk driver
>>> c9194fb623b0 clk: samsung: Remove support for obsolete Exynos4212 CPU clock
>>> d5cd103b06f9 clk: samsung: Remove support for Exynos4212 SoCs in
>>> Exynos CLKOUT driver
>>> 7c4f63ec94a1 clk: sunxi-ng: a83t: Fix invalid csi-mclk mux offset
>>> WORKS
>>> 0a4e632b6f9d Merge branch 'clk-fixes' into clk-next
>>>
>>> BR,
>>> CK
>>>
>>>
>>> On 22 May 2018 at 21:27, Code Kipper <codekipper-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
>>>> Hi Markus,
>>>> I've been able to reproduce this on my A31 board with linux-next(plays
>>>> at 44100 but not at 96K)so it is an issue. I also tested the same code
>>>> on my A64 and it worked fine, The blocks are very similar so it maybe
>>>> a clocking issue, I'll look into this.
>>>> BR,
>>>> CK
>>>>
>>>> On 21 May 2018 at 20:17, Markus Mitsch <markusmitsch86-gM/Ye1E23mwN+BqQ9rBEUg@public.gmane.org> wrote:
>>>>> 2018-05-21 15:17 GMT+02:00 Maxime Ripard <maxime.ripard-LDxbnhwyfcJBDgjK7y7TUQ@public.gmane.org>:
>>>>>> On Mon, May 21, 2018 at 02:43:55PM +0200, Markus Mitsch wrote:
>>>>>>> 2018-05-21 11:43 GMT+02:00 Markus Mitsch <markusmitsch86-gM/Ye1E23mwN+BqQ9rBEUg@public.gmane.org>:
>>>>>>> >>> Hello,
>>>>>>> >>>
>>>>>>> >>> i have a new problem. if i want to output sound from optical i get
>>>>>>> >>> distorted sound. At 44.1 khz its like an old vinly, but at 48 khz and
>>>>>>> >>> higher it is very bad. thanks in advance for your time.
>>>>>>> >
>>>>>>> >> Hi Markus,
>>>>>>> >> I haven't tested it for a while but when I did it worked as high as
>>>>>>> >> 96K with no issues. Does it work ok with the original cubietruck
>>>>>>> >> software?,
>>>>>>> >> CK
>>>>>>> >
>>>>>>> > Hello,
>>>>>>> > i am using arch linux with mainline kernel 4.16.8. parhaps i can try
>>>>>>> > today with official image.
>>>>>>> > mfg
>>>>>>>
>>>>>>> Hi.
>>>>>>> So, with debian image from 2014 with 3.4.79 kernel the sound is clear
>>>>>>> up to 96 khz.
>>>>>>
>>>>>> You said that it is a new problem, when was it introduced? Could you
>>>>>> bisect the issue?
>>>>>>
>>>>>> Maxime
>>>>>
>>>>> Hello.
>>>>> I dont know since when this problem appears. When saying "new" i
>>>>> refered to my older post. I tried to look into the dts file and the
>>>>> code
>>>>> but i am afraid this is a bit too high for me. I can say that the
>>>>> analog output works, even at high sampling rates.
>>>>> And I have read something about a clock source could be the problem.
>>>>> I'm sorry I can't tell you any more.
>>>>>
>>>>> mfg
>>>>> Markus
>>
>> Hello,
>> i tried undoing part of the commit
>> de3448519194 clk: sunxi-ng: sun4i: Use sigma-delta modulation for audio PLL
>> by setting the value for the divider from 1 to 4 again on both places.
>> with no success.
>>
>> markus
>
> i now had a little success. the output works!!
> i disabled the sigma-delta support by making
> ccu_sdm_helper_enable in drivers/clk/sunxi-ng/ccu_sdm.c  return imediately.
>
> markus

Re: [BUG] On Allwinner A20 Board Cubietruck distorted SPDIF opical output

[Chen-Yu Tsai]

Hi,

On Sun, Jun 10, 2018 at 1:46 AM, Markus Mitsch
<markusmitsch86-gM/Ye1E23mwN+BqQ9rBEUg@public.gmane.org> wrote:
> hi.
>
> i posted my answer wrong so here is my finding:
> i made the output work by disabling the sigma delta modulation.
> after compiling the kernel  i can output through spdif with any rate.

So it seems that the default mid-level bias current setting for PLL2
interferes with sigma-delta modulation. Similarly, the VCO bias setting
also interferes.

Try zeroing out bits 16 ~ 25 in the PLL2 tuning register (0x01c2000c)
with sigma delta on, and see if that fixes things.

I'll send out a patch later.

ChenYu

> greetings
> markus
>
> 2018-05-26 18:38 GMT+02:00 Markus Mitsch <markusmitsch86-gM/Ye1E23mwN+BqQ9rBEUg@public.gmane.org>:
>> 2018-05-26 16:57 GMT+02:00 Markus Mitsch <markusmitsch86-gM/Ye1E23mwN+BqQ9rBEUg@public.gmane.org>:
>>> 2018-05-23 17:12 GMT+02:00 Code Kipper <codekipper-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>:
>>>> Hi All,
>>>> I can see that the problem was introduced somewhere here
>>>>
>>>> DOESN'T WORK
>>>> ffc3eb6f3a83 Merge tag 'sunxi-clk-for-4.15' of
>>>> https://git.kernel.org/pub/scm/linux/kernel/git/sunxi/linux into
>>>> clk-next
>>>> ae74ac082886 Merge tag 'clk-v4.15-exynos-pm' of
>>>> git://git.kernel.org/pub/scm/linux/kernel/git/snawrocki/clk into
>>>> clk-next
>>>> faa865f18cb7 Merge tag 'clk-v4.15-samsung' of
>>>> git://git.kernel.org/pub/scm/linux/kernel/git/snawrocki/clk into
>>>> clk-next
>>>> f09a6b86fdae Merge tag 'clk-renesas-for-v4.15-tag1' of
>>>> git://git.kernel.org/pub/scm/linux/kernel/git/geert/renesas-drivers
>>>> into clk-next
>>>> 4328a2186e51 clk: sunxi-ng: sun4i: Export video PLLs
>>>> 553c7d5ba2fe clk: sunxi-ng: Add A83T display clocks
>>>> 7679eb20353d clk: samsung: Add a separate driver for Exynos4412 ISP clocks
>>>> 8ca8ac102484 clk: samsung: Add dt bindings for Exynos4412 ISP clock controller
>>>> 75920aac275a clk: samsung: Instantiate Exynos4412 ISP clocks only when available
>>>> 24ea78a09f89 clk: sunxi-ng: sun8i: a23: Use sigma-delta modulation for audio PLL
>>>> ee6501d69217 clk: sunxi-ng: sun6i: Use sigma-delta modulation for audio PLL
>>>> 042f7f8f9715 clk: sunxi-ng: sun5i: Use sigma-delta modulation for audio PLL
>>>> de3448519194 clk: sunxi-ng: sun4i: Use sigma-delta modulation for audio PLL
>>>> a5e3e2b2ef85 clk: sunxi-ng: sun8i: h3: Use sigma-delta modulation for audio PLL
>>>> 392ba5fafcdf clk: sunxi-ng: nm: Add support for sigma-delta modulation
>>>> 05d2eaac96d4 clk: sunxi-ng: Add sigma-delta modulation support
>>>> 4cdbc40d64d4 clk: sunxi-ng: nm: Check if requested rate is supported
>>>> by fractional clock
>>>> d51fe3ba9773 clk: sunxi-ng: sun5i: Fix bit offset of audio PLL post-divider
>>>> 4e8975cbb516 clk: samsung: exynos5433: mark PM functions as __maybe_unused
>>>> a4f21e9ceb5c clk: samsung: Remove obsolete clkdev alias support
>>>> 45d882daf88a clk: samsung: Add explicit MPLL, EPLL clkdev aliases in
>>>> S3C2443 driver
>>>> cd05417f728b clk: samsung: Rework clkdev alias handling in S3C2443 driver
>>>> efea8d377184 clk: samsung: Rework clkdev alias handling in Exynos5440 driver
>>>> 36ba48240b19 clk: samsung: Drop useless alias in Exynos5420 clk driver
>>>> 29964890f31c clk: samsung: Remove clkdev alias support in Exynos5250 clk driver
>>>> 6de08891c896 clk: samsung: Remove double assignment of CLK_ARM_CLK in
>>>> Exynos4 driver
>>>> 58f4a5ff3a62 clk: samsung: Remove clkdev alias support in Exynos4 clk driver
>>>> c9194fb623b0 clk: samsung: Remove support for obsolete Exynos4212 CPU clock
>>>> d5cd103b06f9 clk: samsung: Remove support for Exynos4212 SoCs in
>>>> Exynos CLKOUT driver
>>>> 7c4f63ec94a1 clk: sunxi-ng: a83t: Fix invalid csi-mclk mux offset
>>>> WORKS
>>>> 0a4e632b6f9d Merge branch 'clk-fixes' into clk-next
>>>>
>>>> BR,
>>>> CK
>>>>
>>>>
>>>> On 22 May 2018 at 21:27, Code Kipper <codekipper-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
>>>>> Hi Markus,
>>>>> I've been able to reproduce this on my A31 board with linux-next(plays
>>>>> at 44100 but not at 96K)so it is an issue. I also tested the same code
>>>>> on my A64 and it worked fine, The blocks are very similar so it maybe
>>>>> a clocking issue, I'll look into this.
>>>>> BR,
>>>>> CK
>>>>>
>>>>> On 21 May 2018 at 20:17, Markus Mitsch <markusmitsch86-gM/Ye1E23mwN+BqQ9rBEUg@public.gmane.org> wrote:
>>>>>> 2018-05-21 15:17 GMT+02:00 Maxime Ripard <maxime.ripard-LDxbnhwyfcJBDgjK7y7TUQ@public.gmane.org>:
>>>>>>> On Mon, May 21, 2018 at 02:43:55PM +0200, Markus Mitsch wrote:
>>>>>>>> 2018-05-21 11:43 GMT+02:00 Markus Mitsch <markusmitsch86-gM/Ye1E23mwN+BqQ9rBEUg@public.gmane.org>:
>>>>>>>> >>> Hello,
>>>>>>>> >>>
>>>>>>>> >>> i have a new problem. if i want to output sound from optical i get
>>>>>>>> >>> distorted sound. At 44.1 khz its like an old vinly, but at 48 khz and
>>>>>>>> >>> higher it is very bad. thanks in advance for your time.
>>>>>>>> >
>>>>>>>> >> Hi Markus,
>>>>>>>> >> I haven't tested it for a while but when I did it worked as high as
>>>>>>>> >> 96K with no issues. Does it work ok with the original cubietruck
>>>>>>>> >> software?,
>>>>>>>> >> CK
>>>>>>>> >
>>>>>>>> > Hello,
>>>>>>>> > i am using arch linux with mainline kernel 4.16.8. parhaps i can try
>>>>>>>> > today with official image.
>>>>>>>> > mfg
>>>>>>>>
>>>>>>>> Hi.
>>>>>>>> So, with debian image from 2014 with 3.4.79 kernel the sound is clear
>>>>>>>> up to 96 khz.
>>>>>>>
>>>>>>> You said that it is a new problem, when was it introduced? Could you
>>>>>>> bisect the issue?
>>>>>>>
>>>>>>> Maxime
>>>>>>
>>>>>> Hello.
>>>>>> I dont know since when this problem appears. When saying "new" i
>>>>>> refered to my older post. I tried to look into the dts file and the
>>>>>> code
>>>>>> but i am afraid this is a bit too high for me. I can say that the
>>>>>> analog output works, even at high sampling rates.
>>>>>> And I have read something about a clock source could be the problem.
>>>>>> I'm sorry I can't tell you any more.
>>>>>>
>>>>>> mfg
>>>>>> Markus
>>>
>>> Hello,
>>> i tried undoing part of the commit
>>> de3448519194 clk: sunxi-ng: sun4i: Use sigma-delta modulation for audio PLL
>>> by setting the value for the divider from 1 to 4 again on both places.
>>> with no success.
>>>
>>> markus
>>
>> i now had a little success. the output works!!
>> i disabled the sigma-delta support by making
>> ccu_sdm_helper_enable in drivers/clk/sunxi-ng/ccu_sdm.c  return imediately.
>>
>> markus

Re: [BUG] On Allwinner A20 Board Cubietruck distorted SPDIF opical output

['Markus Mitsch' via linux-sunxi]

Hello,
sorry for the late answer. all my problems with optical out seem to
be solved with this. i still dont understand why without sigma delta
modulation the 44.1 khz did not work while all other sampling rates
were ok.
but now i can leave sdm on an did not see a sampling rate fail. thanks to all.

greetings
markus
Am Mi., 8. Aug. 2018 um 11:59 Uhr schrieb Chen-Yu Tsai <wens-jdAy2FN1RRM@public.gmane.org>:
>
> Hi,
>
> On Sun, Jun 10, 2018 at 1:46 AM, Markus Mitsch
> <markusmitsch86-gM/Ye1E23mwN+BqQ9rBEUg@public.gmane.org> wrote:
> > hi.
> >
> > i posted my answer wrong so here is my finding:
> > i made the output work by disabling the sigma delta modulation.
> > after compiling the kernel  i can output through spdif with any rate.
>
> So it seems that the default mid-level bias current setting for PLL2
> interferes with sigma-delta modulation. Similarly, the VCO bias setting
> also interferes.
>
> Try zeroing out bits 16 ~ 25 in the PLL2 tuning register (0x01c2000c)
> with sigma delta on, and see if that fixes things.
>
> I'll send out a patch later.
>
> ChenYu
>
> > greetings
> > markus
> >
> > 2018-05-26 18:38 GMT+02:00 Markus Mitsch <markusmitsch86-gM/Ye1E23mwN+BqQ9rBEUg@public.gmane.org>:
> >> 2018-05-26 16:57 GMT+02:00 Markus Mitsch <markusmitsch86-gM/Ye1E23mwN+BqQ9rBEUg@public.gmane.org>:
> >>> 2018-05-23 17:12 GMT+02:00 Code Kipper <codekipper-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>:
> >>>> Hi All,
> >>>> I can see that the problem was introduced somewhere here
> >>>>
> >>>> DOESN'T WORK
> >>>> ffc3eb6f3a83 Merge tag 'sunxi-clk-for-4.15' of
> >>>> https://git.kernel.org/pub/scm/linux/kernel/git/sunxi/linux into
> >>>> clk-next
> >>>> ae74ac082886 Merge tag 'clk-v4.15-exynos-pm' of
> >>>> git://git.kernel.org/pub/scm/linux/kernel/git/snawrocki/clk into
> >>>> clk-next
> >>>> faa865f18cb7 Merge tag 'clk-v4.15-samsung' of
> >>>> git://git.kernel.org/pub/scm/linux/kernel/git/snawrocki/clk into
> >>>> clk-next
> >>>> f09a6b86fdae Merge tag 'clk-renesas-for-v4.15-tag1' of
> >>>> git://git.kernel.org/pub/scm/linux/kernel/git/geert/renesas-drivers
> >>>> into clk-next
> >>>> 4328a2186e51 clk: sunxi-ng: sun4i: Export video PLLs
> >>>> 553c7d5ba2fe clk: sunxi-ng: Add A83T display clocks
> >>>> 7679eb20353d clk: samsung: Add a separate driver for Exynos4412 ISP clocks
> >>>> 8ca8ac102484 clk: samsung: Add dt bindings for Exynos4412 ISP clock controller
> >>>> 75920aac275a clk: samsung: Instantiate Exynos4412 ISP clocks only when available
> >>>> 24ea78a09f89 clk: sunxi-ng: sun8i: a23: Use sigma-delta modulation for audio PLL
> >>>> ee6501d69217 clk: sunxi-ng: sun6i: Use sigma-delta modulation for audio PLL
> >>>> 042f7f8f9715 clk: sunxi-ng: sun5i: Use sigma-delta modulation for audio PLL
> >>>> de3448519194 clk: sunxi-ng: sun4i: Use sigma-delta modulation for audio PLL
> >>>> a5e3e2b2ef85 clk: sunxi-ng: sun8i: h3: Use sigma-delta modulation for audio PLL
> >>>> 392ba5fafcdf clk: sunxi-ng: nm: Add support for sigma-delta modulation
> >>>> 05d2eaac96d4 clk: sunxi-ng: Add sigma-delta modulation support
> >>>> 4cdbc40d64d4 clk: sunxi-ng: nm: Check if requested rate is supported
> >>>> by fractional clock
> >>>> d51fe3ba9773 clk: sunxi-ng: sun5i: Fix bit offset of audio PLL post-divider
> >>>> 4e8975cbb516 clk: samsung: exynos5433: mark PM functions as __maybe_unused
> >>>> a4f21e9ceb5c clk: samsung: Remove obsolete clkdev alias support
> >>>> 45d882daf88a clk: samsung: Add explicit MPLL, EPLL clkdev aliases in
> >>>> S3C2443 driver
> >>>> cd05417f728b clk: samsung: Rework clkdev alias handling in S3C2443 driver
> >>>> efea8d377184 clk: samsung: Rework clkdev alias handling in Exynos5440 driver
> >>>> 36ba48240b19 clk: samsung: Drop useless alias in Exynos5420 clk driver
> >>>> 29964890f31c clk: samsung: Remove clkdev alias support in Exynos5250 clk driver
> >>>> 6de08891c896 clk: samsung: Remove double assignment of CLK_ARM_CLK in
> >>>> Exynos4 driver
> >>>> 58f4a5ff3a62 clk: samsung: Remove clkdev alias support in Exynos4 clk driver
> >>>> c9194fb623b0 clk: samsung: Remove support for obsolete Exynos4212 CPU clock
> >>>> d5cd103b06f9 clk: samsung: Remove support for Exynos4212 SoCs in
> >>>> Exynos CLKOUT driver
> >>>> 7c4f63ec94a1 clk: sunxi-ng: a83t: Fix invalid csi-mclk mux offset
> >>>> WORKS
> >>>> 0a4e632b6f9d Merge branch 'clk-fixes' into clk-next
> >>>>
> >>>> BR,
> >>>> CK
> >>>>
> >>>>
> >>>> On 22 May 2018 at 21:27, Code Kipper <codekipper-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> >>>>> Hi Markus,
> >>>>> I've been able to reproduce this on my A31 board with linux-next(plays
> >>>>> at 44100 but not at 96K)so it is an issue. I also tested the same code
> >>>>> on my A64 and it worked fine, The blocks are very similar so it maybe
> >>>>> a clocking issue, I'll look into this.
> >>>>> BR,
> >>>>> CK
> >>>>>
> >>>>> On 21 May 2018 at 20:17, Markus Mitsch <markusmitsch86-gM/Ye1E23mwN+BqQ9rBEUg@public.gmane.org> wrote:
> >>>>>> 2018-05-21 15:17 GMT+02:00 Maxime Ripard <maxime.ripard-LDxbnhwyfcJBDgjK7y7TUQ@public.gmane.org>:
> >>>>>>> On Mon, May 21, 2018 at 02:43:55PM +0200, Markus Mitsch wrote:
> >>>>>>>> 2018-05-21 11:43 GMT+02:00 Markus Mitsch <markusmitsch86-gM/Ye1E23mwN+BqQ9rBEUg@public.gmane.org>:
> >>>>>>>> >>> Hello,
> >>>>>>>> >>>
> >>>>>>>> >>> i have a new problem. if i want to output sound from optical i get
> >>>>>>>> >>> distorted sound. At 44.1 khz its like an old vinly, but at 48 khz and
> >>>>>>>> >>> higher it is very bad. thanks in advance for your time.
> >>>>>>>> >
> >>>>>>>> >> Hi Markus,
> >>>>>>>> >> I haven't tested it for a while but when I did it worked as high as
> >>>>>>>> >> 96K with no issues. Does it work ok with the original cubietruck
> >>>>>>>> >> software?,
> >>>>>>>> >> CK
> >>>>>>>> >
> >>>>>>>> > Hello,
> >>>>>>>> > i am using arch linux with mainline kernel 4.16.8. parhaps i can try
> >>>>>>>> > today with official image.
> >>>>>>>> > mfg
> >>>>>>>>
> >>>>>>>> Hi.
> >>>>>>>> So, with debian image from 2014 with 3.4.79 kernel the sound is clear
> >>>>>>>> up to 96 khz.
> >>>>>>>
> >>>>>>> You said that it is a new problem, when was it introduced? Could you
> >>>>>>> bisect the issue?
> >>>>>>>
> >>>>>>> Maxime
> >>>>>>
> >>>>>> Hello.
> >>>>>> I dont know since when this problem appears. When saying "new" i
> >>>>>> refered to my older post. I tried to look into the dts file and the
> >>>>>> code
> >>>>>> but i am afraid this is a bit too high for me. I can say that the
> >>>>>> analog output works, even at high sampling rates.
> >>>>>> And I have read something about a clock source could be the problem.
> >>>>>> I'm sorry I can't tell you any more.
> >>>>>>
> >>>>>> mfg
> >>>>>> Markus
> >>>
> >>> Hello,
> >>> i tried undoing part of the commit
> >>> de3448519194 clk: sunxi-ng: sun4i: Use sigma-delta modulation for audio PLL
> >>> by setting the value for the divider from 1 to 4 again on both places.
> >>> with no success.
> >>>
> >>> markus
> >>
> >> i now had a little success. the output works!!
> >> i disabled the sigma-delta support by making
> >> ccu_sdm_helper_enable in drivers/clk/sunxi-ng/ccu_sdm.c  return imediately.
> >>
> >> markus