Re: audio usemixer breaks vs(4) Re: CVS commit: src/sys/dev

[Tetsuya Isaki]

At Wed, 8 Nov 2017 10:10:52 +1100,
Nathanial Sloss wrote:
> > With this change, vs(4) no longer works even if usemixer=0.
> > Please revert it, and don't break without public discussion.
> 
> vs audio with the mixer enabled has been fixed in a follow up commit.

Thank you for fixing.  But if you knew it would break something,
you should have asked it before commiting.

> The change was intended to benifit slow (antiquated) computers that may have 
> trouble with the extra code in the audio path with mixing enabled.

You should have discussed it before commiting.  There might
have been another solution or strategy.

> The reason as to why vs audio does not work with usemixer=0 is that you 
> decided to remove what you considered to be dead code (mulaw, alaw, unsigned 
> linear filters).

No, it was certainly a dead code at that time.  You had to
explain/discuss about such MI-MD destructive changes before you
did it.

In the old days, the MI-MD specification of set_params is
``mulaw is mandatory, others are optional.''
Last year, you imported in-kernel mixing and changed the spec to
``slinearNN_ is mandatory, others are never used''
without any explanation/discussion.  It broke some minor devices
which did not support slinear. [*]
Half a year ago, I noticed it and repaired vs(4) [**].  And I
deleted the code which is no longer used by your silent change.

Two days ago, you commited usemixer feature.  You changed the spec
to ``slinearNN_ is mandatory, others are optional''
without any explanation/discussion.
And you says that the reason was my deletion.

 *: At least one driver is still broken by this change.  Although
there may be no users for the device :-)
 **: There was another MI-MD destructive change about slinear8.

> The time I've spent on audio/ossaudio/bluetooth audio has been at least two 
> years non-stop and I will devote more time to it in future - so I think it's 
> fair to say we've all spent a lot of time perfecting NetBSD.

You used your two years for youself.
I have not used this half a year for myself.

> Best regards and I hope to collaborate with you on audio in future as it is a 
> mutual interest.

Please go ahead.
I have no opinion anymore.
---
Tetsuya Isaki 

Re: audio usemixer breaks vs(4) Re: CVS commit: src/sys/dev

[Pierre Pronchery]

Hi Nathanial, Tetsuya, (Christos)

this is one more thank you to Nathanial for this awesome work.

He's going to great length to make sure his changes work everywhere
every time, and even then it is very, very difficult to achieve a net
positive on every platform in every commit - we are supporting so many!

Let's just be a bit nicer to each other, fix issues as we see them
appear, and keep up the great work.

Cheers,
-- khorben

On 08/11/2017 00:10, Nathanial Sloss wrote:
> Hello Tetsuya and Christos,
> 
> 
> On Tue, 7 Nov 2017 14:45:47 Tetsuya Isaki wrote:
>> nat@, (cc: christos@)
>>
>> With this change, vs(4) no longer works even if usemixer=0.
>> Please revert it, and don't break without public discussion.
> 
> vs audio with the mixer enabled has been fixed in a follow up commit.
> 
> The change was intended to benifit slow (antiquated) computers that may have 
> trouble with the extra code in the audio path with mixing enabled.
> 
>>
>> This is the 3rd time you broke vs(4).
>> You broke and I repaired.
>> You broke and I repaired.
>> I spent a lot of time this half year.
> 
> The reason as to why vs audio does not work with usemixer=0 is that you 
> decided to remove what you considered to be dead code (mulaw, alaw, unsigned 
> linear filters).
> 
> The time I've spent on audio/ossaudio/bluetooth audio has been at least two 
> years non-stop and I will devote more time to it in future - so I think it's 
> fair to say we've all spent a lot of time perfecting NetBSD.
> 
>>
>> If you can not consider about supported devices (at least
>> devices which has active users), please don't touch audio.
>>
> 
> Every time I make a change I test it with emulations of various computers, 
> usb 
> audio devices, all of the computers I own and bluetooth.
> 
> Also an x68k emulation since x68k has been a problem with the changes.
> 
> I think to hold the merits of what I done to audio to x68k's abilities to be 
> unfair.
> 
> I'm actively trying to innovate and to bring NetBSD's audio server into the 
> 21 
> century whist still providing a solution for legacy systems we support.
> 
> I hope to contine to work on audio in future - if there is an emmense 
> objection from others I will stop audio completely and work on something else.
> 
> NB: Since I've started to alter audio with regards to mixing and such, I've 
> received at least 30 thank you emails from strangers and NetBSD users -  The 
> people we write NetBSD for.
> 
> 
>> Thanks,
> 
>> ---
>> Tetsuya Isaki 
> 
> Best regards and I hope to collaborate with you on audio in future as it is a 
> mutual interest.
> 
> Nat
> 
>>
>> At Tue, 7 Nov 2017 01:13:19 +,
>>
>>> Module Name:src
>>> Committed By:   nat
>>> Date:   Tue Nov  7 01:13:19 UTC 2017
>>>
>>> Modified Files:
>>> src/sys/dev: audio.c audiovar.h
>>>
>>> Log Message:
>>> A sysctl is now available to disable the in kernel mixer.
>>>
>>> sysctl -w hw.hdafg0.usemixer=0
>>>
>>> There currently is a problem draining the last block with the mixer
>>> disabled.  I will fix this in a follow up commit.
>>>
>>> AFAIK there will be a problem wiht vs(4) on x68k with the mixer disabled
>>> as the filters for mulaw, alaw and unsigned linear have been removed post
>>> audio mixing changes.
>>>
>>> Documentation for this sysctl variable will be made to audio.4 in a
>>> follow up commit.
>>>
>>> Ok christos@.
>>>
>>>
>>> To generate a diff of this commit:
>>> cvs rdiff -u -r1.429 -r1.430 src/sys/dev/audio.c
>>> cvs rdiff -u -r1.66 -r1.67 src/sys/dev/audiovar
-- 
khorben



signature.asc
Description: OpenPGP digital signature

Re: audio usemixer breaks vs(4) Re: CVS commit: src/sys/dev

[Nathanial Sloss]

Hello Tetsuya and Christos,


On Tue, 7 Nov 2017 14:45:47 Tetsuya Isaki wrote:
> nat@, (cc: christos@)
> 
> With this change, vs(4) no longer works even if usemixer=0.
> Please revert it, and don't break without public discussion.

vs audio with the mixer enabled has been fixed in a follow up commit.

The change was intended to benifit slow (antiquated) computers that may have 
trouble with the extra code in the audio path with mixing enabled.

> 
> This is the 3rd time you broke vs(4).
> You broke and I repaired.
> You broke and I repaired.
> I spent a lot of time this half year.

The reason as to why vs audio does not work with usemixer=0 is that you 
decided to remove what you considered to be dead code (mulaw, alaw, unsigned 
linear filters).

The time I've spent on audio/ossaudio/bluetooth audio has been at least two 
years non-stop and I will devote more time to it in future - so I think it's 
fair to say we've all spent a lot of time perfecting NetBSD.

> 
> If you can not consider about supported devices (at least
> devices which has active users), please don't touch audio.
> 

Every time I make a change I test it with emulations of various computers, usb 
audio devices, all of the computers I own and bluetooth.

Also an x68k emulation since x68k has been a problem with the changes.

I think to hold the merits of what I done to audio to x68k's abilities to be 
unfair.

I'm actively trying to innovate and to bring NetBSD's audio server into the 21 
century whist still providing a solution for legacy systems we support.

I hope to contine to work on audio in future - if there is an emmense 
objection from others I will stop audio completely and work on something else.

NB: Since I've started to alter audio with regards to mixing and such, I've 
received at least 30 thank you emails from strangers and NetBSD users -  The 
people we write NetBSD for.


> Thanks,

> ---
> Tetsuya Isaki 

Best regards and I hope to collaborate with you on audio in future as it is a 
mutual interest.

Nat

> 
> At Tue, 7 Nov 2017 01:13:19 +,
> 
> > Module Name:src
> > Committed By:   nat
> > Date:   Tue Nov  7 01:13:19 UTC 2017
> > 
> > Modified Files:
> > src/sys/dev: audio.c audiovar.h
> > 
> > Log Message:
> > A sysctl is now available to disable the in kernel mixer.
> > 
> > sysctl -w hw.hdafg0.usemixer=0
> > 
> > There currently is a problem draining the last block with the mixer
> > disabled.  I will fix this in a follow up commit.
> > 
> > AFAIK there will be a problem wiht vs(4) on x68k with the mixer disabled
> > as the filters for mulaw, alaw and unsigned linear have been removed post
> > audio mixing changes.
> > 
> > Documentation for this sysctl variable will be made to audio.4 in a
> > follow up commit.
> > 
> > Ok christos@.
> > 
> > 
> > To generate a diff of this commit:
> > cvs rdiff -u -r1.429 -r1.430 src/sys/dev/audio.c
> > cvs rdiff -u -r1.66 -r1.67 src/sys/dev/audiovar.h

Re: audio usemixer breaks vs(4) Re: CVS commit: src/sys/dev

[Tetsuya Isaki]

At Tue, 07 Nov 2017 12:45:47 +0900,
Tetsuya Isaki wrote:
> 
> nat@, (cc: christos@)
> 
> With this change, vs(4) no longer works even if usemixer=0.
  ^^
Oops, it should be 'usemixer=1'.

If usemixer=1, it plays noise.
If usemixer=0, failed to open sound device.

Anyway, please revert it.
---
Tetsuya Isaki 

> Please revert it, and don't break without public discussion.
> 
> This is the 3rd time you broke vs(4).
> You broke and I repaired.
> You broke and I repaired.
> I spent a lot of time this half year.
> 
> If you can not consider about supported devices (at least
> devices which has active users), please don't touch audio.
> 
> Thanks,
> ---
> Tetsuya Isaki 
> 
> At Tue, 7 Nov 2017 01:13:19 +,
> > Module Name:src
> > Committed By:   nat
> > Date:   Tue Nov  7 01:13:19 UTC 2017
> > 
> > Modified Files:
> > src/sys/dev: audio.c audiovar.h
> > 
> > Log Message:
> > A sysctl is now available to disable the in kernel mixer.
> > sysctl -w hw.hdafg0.usemixer=0
> > 
> > There currently is a problem draining the last block with the mixer
> > disabled.  I will fix this in a follow up commit.
> > 
> > AFAIK there will be a problem wiht vs(4) on x68k with the mixer disabled
> > as the filters for mulaw, alaw and unsigned linear have been removed post
> > audio mixing changes.
> > 
> > Documentation for this sysctl variable will be made to audio.4 in a follow
> > up commit.
> > 
> > Ok christos@.
> > 
> > 
> > To generate a diff of this commit:
> > cvs rdiff -u -r1.429 -r1.430 src/sys/dev/audio.c
> > cvs rdiff -u -r1.66 -r1.67 src/sys/dev/audiovar.h
> 

audio usemixer breaks vs(4) Re: CVS commit: src/sys/dev

[Tetsuya Isaki]

nat@, (cc: christos@)

With this change, vs(4) no longer works even if usemixer=0.
Please revert it, and don't break without public discussion.

This is the 3rd time you broke vs(4).
You broke and I repaired.
You broke and I repaired.
I spent a lot of time this half year.

If you can not consider about supported devices (at least
devices which has active users), please don't touch audio.

Thanks,
---
Tetsuya Isaki 

At Tue, 7 Nov 2017 01:13:19 +,
> Module Name:  src
> Committed By: nat
> Date: Tue Nov  7 01:13:19 UTC 2017
> 
> Modified Files:
>   src/sys/dev: audio.c audiovar.h
> 
> Log Message:
> A sysctl is now available to disable the in kernel mixer.
>   sysctl -w hw.hdafg0.usemixer=0
> 
> There currently is a problem draining the last block with the mixer
> disabled.  I will fix this in a follow up commit.
> 
> AFAIK there will be a problem wiht vs(4) on x68k with the mixer disabled
> as the filters for mulaw, alaw and unsigned linear have been removed post
> audio mixing changes.
> 
> Documentation for this sysctl variable will be made to audio.4 in a follow
> up commit.
> 
> Ok christos@.
> 
> 
> To generate a diff of this commit:
> cvs rdiff -u -r1.429 -r1.430 src/sys/dev/audio.c
> cvs rdiff -u -r1.66 -r1.67 src/sys/dev/audiovar.h