Author Topic: Updated Uniaud32 build  (Read 397475 times)

Mark Szkolnicki

  • Sr. Member
  • ****
  • Posts: 433
  • Karma: +18/-0
    • View Profile
Re: Updated Uniaud32 build
« Reply #855 on: September 19, 2022, 08:18:57 pm »
Heh Paul!


This is VERY helpful! I dug out an old drive with an older copy of my eCS-udrive.vdi and found the source folders for:
uniaud32-linux-3.8.13
uniaud32-linux-3.9.11

A diff of the changes is at https://smedley.id.au/tmp/uniaud32-3.8.13-3.9.11.diff

Unfortunately, this is a large-ish diff - I'll have a look for anything obvious when I get a moment - hopefully it will be something obvious...

Edit: 1st attempt - hda_intel.c from 3.9.11 backlevelled to the 3.8.13 version - https://smedley.id.au/tmp/uniaud32-linux-3.9.11-20220919.zip
Cheers,

Paul

Tried your edit of 3.9.11, and no problems loading - but no sound generated.

New Testlog attached and Best!

M
Vincit Que Se Vincit - "He Who Conquers Self Succeeds"

Mark Szkolnicki

  • Sr. Member
  • ****
  • Posts: 433
  • Karma: +18/-0
    • View Profile
Re: Updated Uniaud32 build
« Reply #856 on: September 19, 2022, 08:44:17 pm »
Hey Mark,

I most definitely take time off :) We don't have skin kids - and just the one dog these days, so my method of relaxation in the evening is to look at code. The funeral is on at a much more civilised time here - started at around 7pm :) But we will record it and watch at our leisure - so we can fast forward the boring bits :) Whilst I was born in the UK ( we emigrated in 1977 when I was two) I wouldn't say I'm a monarchist - but at the same time - why change a system of government that isn't broken?

Anyway I digress - alcohol free day for us today - will have a glass with dinner tomorrow :P

Cheers,

Paul

Actually we found it quite an interesting show of pagentry and tradition.

I was particularly impressed as to those 8 strapping lads from the Royal Grenadiers hoisting that coffin 5 times, sometimes standing in place for minutes at a time, or under slow march for minutes - it was mentioning here that the Queens Coffin is made of oak ......... and is totally lead lined, as she will be in one of the Vaults under St. Georges Cathedral, in I assume a niche beside Philip.

Moving that coffin with all that weight on level ground is bad enough, manhandling wise - moving it up three steep flights of stairs on your shoulders at St. Georges takes a LOT of ultra coordinated effort and immense strength.

Also interesting as to how they fastened the Crown, Sceptre and Orb to the top - there must have been a plate fastened under the Royal Ensign, as at least two of the objects had pipes or spikes fastened to the bottom, which fitted into niches - the Sceptre was held in place by a bracket. I doubt they wanted to drill holes into the top of the coffin ..........

And finally, I didn't envy those soldiers who were required, by tradition to carry their weapons backwards as a sign of respect - to first of all have to slow march an hour and twenty minutes, then tie yourself up in knots at the same time, is something to behold - I suspect a lot of them will be glad to return to operational duty after today .......... !

But we've never seen anything like that in most of our lifetimes, Sir - Anne was born the year Elizabeth and Philip married ............ !

So a stately village style funeral, people behaving themselves in silence most of the time (except for the occasional cheer and throwing flowers) and a fitting sendoff to a long life well lived - and now with yet another sense of permanency gone ........... what's next?

Best!

M
« Last Edit: September 20, 2022, 09:18:11 pm by Mark Szkolnicki »
Vincit Que Se Vincit - "He Who Conquers Self Succeeds"

Dariusz Piatkowski

  • Hero Member
  • *****
  • Posts: 1323
  • Karma: +26/-0
    • View Profile
Re: Updated Uniaud32 build
« Reply #857 on: September 19, 2022, 08:49:35 pm »
Hi Mark,

Long time no talk. Hows it going?

Things are alright here, busy and slowing down at the same time LOL. Both kids are away for University studies, so home is a little "slower"...but even with fewer people around it just seems like we have more stuff to chase after...ugh??? LOL

...Actually, the problem we're trying to solve is with the RealTek ALC892 chipset in one of my base metal systems - something seems to have broken way back when, with Paul's drivers - between the official release of ArcaOS 5.0.6 (based on old ALSA code) and 5.0.7 (where the UniAud 32 is based on Linux 5.14.7 - release v3.01.02)...

Yup, understood that and given that my chipset here is the ATI SB using ALC889 I figured there may be an odd-chance that the two are linked:

Code: [Select]
HARDWARE INFO ('unimix -card')
Command line mixer for UNIAUD32. Version 1.1
Detected UNIAUD version 2.02.05
Detected 1 audio adapter(s)
Card info for adapter 0:
  num: 0
  id: SB
  driver: HDA-Intel
  name: HDA ATI SB
  longname: HDA ATI SB at 0xfe3f8000 irq 16
  mixer: Realtek ALC889
  componenets: HDA:10ec0889,14627640,00100004

Paul Smedley

  • Hero Member
  • *****
  • Posts: 2095
  • Karma: +160/-0
    • View Profile
Re: Updated Uniaud32 build
« Reply #858 on: September 20, 2022, 12:17:56 am »
Hey Mark,

This is VERY helpful! I dug out an old drive with an older copy of my eCS-udrive.vdi and found the source folders for:
uniaud32-linux-3.8.13
uniaud32-linux-3.9.11

A diff of the changes is at https://smedley.id.au/tmp/uniaud32-3.8.13-3.9.11.diff

Unfortunately, this is a large-ish diff - I'll have a look for anything obvious when I get a moment - hopefully it will be something obvious...

Edit: 1st attempt - hda_intel.c from 3.9.11 backlevelled to the 3.8.13 version - https://smedley.id.au/tmp/uniaud32-linux-3.9.11-20220919.zip
Cheers,

Tried your edit of 3.9.11, and no problems loading - but no sound generated.


:) Not surprised - swapping out that source file only was a bit of a WAG... Tonight I'll aim to build the following:
1) ALSA 3.8.13 core code; with hda code from 3.9.11
2) ALSA 3.9.11 core code; with hda code from 3.8.13

One of those should still produce sound - which should quickly split whether the issue is in the ALSA 3.9.11 core code; or the hda code.

More tonight...

Cheers,

Paul.

Paul Smedley

  • Hero Member
  • *****
  • Posts: 2095
  • Karma: +160/-0
    • View Profile
Re: Updated Uniaud32 build
« Reply #859 on: September 20, 2022, 10:44:47 am »
Me again,


This is VERY helpful! I dug out an old drive with an older copy of my eCS-udrive.vdi and found the source folders for:
uniaud32-linux-3.8.13
uniaud32-linux-3.9.11

A diff of the changes is at https://smedley.id.au/tmp/uniaud32-3.8.13-3.9.11.diff

Unfortunately, this is a large-ish diff - I'll have a look for anything obvious when I get a moment - hopefully it will be something obvious...

Edit: 1st attempt - hda_intel.c from 3.9.11 backlevelled to the 3.8.13 version - https://smedley.id.au/tmp/uniaud32-linux-3.9.11-20220919.zip
Cheers,

Tried your edit of 3.9.11, and no problems loading - but no sound generated.


:) Not surprised - swapping out that source file only was a bit of a WAG... Tonight I'll aim to build the following:
1) ALSA 3.8.13 core code; with hda code from 3.9.11
2) ALSA 3.9.11 core code; with hda code from 3.8.13

One of those should still produce sound - which should quickly split whether the issue is in the ALSA 3.9.11 core code; or the hda code.

OK....
1) https://smedley.id.au/tmp/uniaud32-linux-3.8.13-20220920.zip <== sound 'core' from 3.8.13; pci/hda code from 3.9.11
2) https://smedley.id.au/tmp/uniaud32-linux-3.9.11-20220920.zip <== sound 'core' from 3.9.11; pci/hda code from 3.8.13

My 'gut' feel is that the problem lies in the hda code, and that 2) above will continue to produce sound.... but we'll see....

I do see one interesting thing in the 3.9.13 log you posted....
Code: [Select]
hda-codec: enabling ADC switching
Not sure if it's part of the problem or not yet though...

Cheers,

Paul

Eugene Tucker

  • Sr. Member
  • ****
  • Posts: 369
  • Karma: +12/-0
    • View Profile
Re: Updated Uniaud32 build
« Reply #860 on: September 20, 2022, 05:35:03 pm »
Neither of the two latest drivers are working on my system. No sound at all in HDA.

mauro

  • Sr. Member
  • ****
  • Posts: 410
  • Karma: +3/-0
    • View Profile
Re: Updated Uniaud32 build
« Reply #861 on: September 20, 2022, 07:37:56 pm »
Hi Mauro

I spruced it up a bit today and have attached it below - hope you will find it of some use.

If anyone has any comments related to the guide, let me know and I will update it.

Best!

Mark

Thank to you as well Mark,  very useful job
[/quote]

Ok Mark, applied suggested editing on config.sys and also downloaded-executed UniaudMix.exe v 0.82 ; result is "no Uniaud device found"  This should have something to do with no uniaud16.sys driver available in the MMOS2 folder.  At this point it is clear I'm missing something very trivial, I feel have done all suggested steps:  downloaded-executed the Uniaud installer, replaced the existing files in MMOS2 with the Paul's updated ones, edited config.sys as suggested in your Testing Guide, but at boot the warning complains that uniaud16.sys is not found as per config.sys call (obvious, I know it is missing).  And of course no sound in the system.
Hope to learn next step from some of you.
Thank you meantime.

Mark Szkolnicki

  • Sr. Member
  • ****
  • Posts: 433
  • Karma: +18/-0
    • View Profile
Re: Updated Uniaud32 build
« Reply #862 on: September 20, 2022, 08:42:07 pm »
Heh Paul!


OK....
1) https://smedley.id.au/tmp/uniaud32-linux-3.8.13-20220920.zip <== sound 'core' from 3.8.13; pci/hda code from 3.9.11
2) https://smedley.id.au/tmp/uniaud32-linux-3.9.11-20220920.zip <== sound 'core' from 3.9.11; pci/hda code from 3.8.13

My 'gut' feel is that the problem lies in the hda code, and that 2) above will continue to produce sound.... but we'll see....

I do see one interesting thing in the 3.9.13 log you posted....
Code: [Select]
hda-codec: enabling ADC switching
Not sure if it's part of the problem or not yet though...

Cheers,

Paul

Well your gut feeling was correct, Sir:

Loading UniAud32.sys 3.9.11 Debug 20220920:

Lovely clear sound produced in both system sounds and PM123 (Played The Poppy Family - Where Evil Grows ......... seemed appropriate  8).)

Loading UniAud32.sys 3.8.13 Debug 20220920:

No sound produced

So it seems to be narrowed down to the pci/hda code ........... !

Testlogs for both attached below.
Vincit Que Se Vincit - "He Who Conquers Self Succeeds"

Mark Szkolnicki

  • Sr. Member
  • ****
  • Posts: 433
  • Karma: +18/-0
    • View Profile
Re: Updated Uniaud32 build
« Reply #863 on: September 20, 2022, 08:59:21 pm »

Things are alright here, busy and slowing down at the same time LOL. Both kids are away for University studies, so home is a little "slower"...but even with fewer people around it just seems like we have more stuff to chase after...ugh??? LOL


Good Job, Sir - Busy and slowing down is exactly how I would describe this Year of Strange - antithetical to the extreme.


Yup, understood that and given that my chipset here is the ATI SB using ALC889 I figured there may be an odd-chance that the two are linked:


Has it ever worked before? And if so, what version of UniAud32 was it?

M
« Last Edit: September 20, 2022, 09:35:12 pm by Mark Szkolnicki »
Vincit Que Se Vincit - "He Who Conquers Self Succeeds"

Mark Szkolnicki

  • Sr. Member
  • ****
  • Posts: 433
  • Karma: +18/-0
    • View Profile
Re: Updated Uniaud32 build
« Reply #864 on: September 20, 2022, 09:16:25 pm »
Heh Mauro!


Ok Mark, applied suggested editing on config.sys and also downloaded-executed UniaudMix.exe v 0.82 ; result is "no Uniaud device found"  This should have something to do with no uniaud16.sys driver available in the MMOS2 folder.  At this point it is clear I'm missing something very trivial, I feel have done all suggested steps:  downloaded-executed the Uniaud installer, replaced the existing files in MMOS2 with the Paul's updated ones, edited config.sys as suggested in your Testing Guide, but at boot the warning complains that uniaud16.sys is not found as per config.sys call (obvious, I know it is missing).  And of course no sound in the system.
Hope to learn next step from some of you.
Thank you meantime.


Paul would know better than I would, but I think having UniAud16 in X:/MMOS2 is still required, in all cases. I have it still in all the partitions I've tested - whatever was installed originally is fine, and to my knowledge never has to change or be removed to use the UniAud32 drivers:

I also always leave the UniAud16.sys statement in place as well, under the UniAud32.sys statement (ie. not removed or REM'd out):

DEVICE=X:\MMOS2\UNIAUD16.SYS /C

The only item really required is changing UniAud32 to the version being tested, in X;\MMOS2 - I usually install all the files Paul provides but don't know if that is really required - testing the legacy UniAud32 files Paul provided at the moment, to solve the ALC892 problem, I have only installed the appropriate UniAud32.sys, and it seems to work - for the newest versions I don't know if .sym and .map are also needed, but never has hurt yet to include them.

I will say that when I've been generating testlogs I always do it with the debug version, as opposed to release.

So I suspect not having UniAud16.sys (and possibly UniAud16.sym as well) or removing the  UniAud16.sys statement in Config.sys, or both, is probably your problem.

Let us know - yet another item possible for v1.2 of the guide, along with Andy Willis's suggestion related to volume.

BTW include the sound chipset if you could as well (testlog would probably show that if you don't know) - I'm currently going through the posts to see if I can compile a list of chipsets that have been tested with the new UniAud drivers, and work when I get some free moments - might be useful to include in the guide in future, or to have Martin post as a Wiki.

Best!

M
« Last Edit: September 20, 2022, 09:41:19 pm by Mark Szkolnicki »
Vincit Que Se Vincit - "He Who Conquers Self Succeeds"

Paul Smedley

  • Hero Member
  • *****
  • Posts: 2095
  • Karma: +160/-0
    • View Profile
Re: Updated Uniaud32 build
« Reply #865 on: September 21, 2022, 12:59:58 am »
Hi Eugene,
Neither of the two latest drivers are working on my system. No sound at all in HDA.

If you're referring to the 0920 builds - I should probably have made it clearer that these are directed towards Mark - and trying to diagnose what change in the linux source code broke sound on his ALC892. They are based on VERY old linux kernels - probably pre-dating the release of the chipsets in your hardware.

Cheers,

Paul.

Paul Smedley

  • Hero Member
  • *****
  • Posts: 2095
  • Karma: +160/-0
    • View Profile
Re: Updated Uniaud32 build
« Reply #866 on: September 21, 2022, 01:02:32 am »
Hey Mark,

OK....
1) https://smedley.id.au/tmp/uniaud32-linux-3.8.13-20220920.zip <== sound 'core' from 3.8.13; pci/hda code from 3.9.11
2) https://smedley.id.au/tmp/uniaud32-linux-3.9.11-20220920.zip <== sound 'core' from 3.9.11; pci/hda code from 3.8.13

My 'gut' feel is that the problem lies in the hda code, and that 2) above will continue to produce sound.... but we'll see....

I do see one interesting thing in the 3.9.13 log you posted....
Code: [Select]
hda-codec: enabling ADC switching
Not sure if it's part of the problem or not yet though...


Well your gut feeling was correct, Sir:

Loading UniAud32.sys 3.9.11 Debug 20220920:

Lovely clear sound produced in both system sounds and PM123 (Played The Poppy Family - Where Evil Grows ......... seemed appropriate  8).)

Loading UniAud32.sys 3.8.13 Debug 20220920:

No sound produced

So it seems to be narrowed down to the pci/hda code ........... !

Testlogs for both attached below.

Thanks - tonight, I'll aim to produce several builds, slowly replacing the 3.8.x pci/hda/* code with the 3.9.x pci/hda/* code - ideally one source file at a time.

Cheers,

Paul.

Paul Smedley

  • Hero Member
  • *****
  • Posts: 2095
  • Karma: +160/-0
    • View Profile
Re: Updated Uniaud32 build
« Reply #867 on: September 21, 2022, 11:50:05 am »
Hi Mark,
Thanks - tonight, I'll aim to produce several builds, slowly replacing the 3.8.x pci/hda/* code with the 3.9.x pci/hda/* code - ideally one source file at a time.

Note: these are specifically directed at trying to understand what broken with his hardware between the sound code in linux 3.8, and linux 3.9.

Code: [Select]
hda_intel.c https://smedley.id.au/tmp/uniaud32-linux-3.9.11-20220921.zip

hda_auto_parser.c https://smedley.id.au/tmp/uniaud32-linux-3.9.11-20220921b.zip
hda_auto_parser.h
hda_codec.c
hda_codec.h
hda_local.h
hda_proc.c
hda_generic.c
hda_generic.h
patch_realtek.c

hda_beep.c https://smedley.id.au/tmp/uniaud32-linux-3.9.11-20220921c.zip
hda_beep.h


hda_eld.c https://smedley.id.au/tmp/uniaud32-linux-3.9.11-20220921d.zip

hda_hwdep.c https://smedley.id.au/tmp/uniaud32-linux-3.9.11-20220921e.zip

hda_jack.c https://smedley.id.au/tmp/uniaud32-linux-3.9.11-20220921f.zip
hda_jack.h



hda_intel_trace.h https://smedley.id.au/tmp/uniaud32-linux-3.9.11-20220921g.zip
hda_patch.h
hda_trace.h

If I had to bet, uniaud32-linux-3.9.11-20220921b.zip is the one that will break things. It's both the most obvious source files impacted, but also it's the largest code change - unfortunately there are interdependencies between those source files, so changing one at a time wasn't possible.

Hopefully I'm wrong, and it will be once of the smaller code changes.

Cheers,

Paul

Mark Szkolnicki

  • Sr. Member
  • ****
  • Posts: 433
  • Karma: +18/-0
    • View Profile
Re: Updated Uniaud32 build
« Reply #868 on: September 21, 2022, 04:56:25 pm »
Heh Paul!

Had our first frost here last night - temps hit -2 C,. so we're now into the major change that is our fall (three weeks ago in Alberta we hit 35 C, and it gets a LOT colder than this in the dead of our winter  :(

Hi Mark,
Thanks - tonight, I'll aim to produce several builds, slowly replacing the 3.8.x pci/hda/* code with the 3.9.x pci/hda/* code - ideally one source file at a time.

Note: these are specifically directed at trying to understand what broken with his hardware between the sound code in linux 3.8, and linux 3.9.

Code: [Select]
hda_intel.c https://smedley.id.au/tmp/uniaud32-linux-3.9.11-20220921.zip

hda_auto_parser.c https://smedley.id.au/tmp/uniaud32-linux-3.9.11-20220921b.zip
hda_auto_parser.h
hda_codec.c
hda_codec.h
hda_local.h
hda_proc.c
hda_generic.c
hda_generic.h
patch_realtek.c

hda_beep.c https://smedley.id.au/tmp/uniaud32-linux-3.9.11-20220921c.zip
hda_beep.h


hda_eld.c https://smedley.id.au/tmp/uniaud32-linux-3.9.11-20220921d.zip

hda_hwdep.c https://smedley.id.au/tmp/uniaud32-linux-3.9.11-20220921e.zip

hda_jack.c https://smedley.id.au/tmp/uniaud32-linux-3.9.11-20220921f.zip
hda_jack.h



hda_intel_trace.h https://smedley.id.au/tmp/uniaud32-linux-3.9.11-20220921g.zip
hda_patch.h
hda_trace.h

If I had to bet, uniaud32-linux-3.9.11-20220921b.zip is the one that will break things. It's both the most obvious source files impacted, but also it's the largest code change - unfortunately there are interdependencies between those source files, so changing one at a time wasn't possible.

Hopefully I'm wrong, and it will be once of the smaller code changes.

Cheers,

Paul

All of them (a to g) on my computer and unzipped - just needed to make sure I got them all - will go methodically from top to bottom to see whats shaking and generate testlogs for each.

Thanks for this tremendous amount of detail and work, Sir - I'm hoping this may solve other peoples problems as well (ie. not just a personal tilt at the windmill) - if other chipsets haven't worked in the past this may be very useful.

M
Vincit Que Se Vincit - "He Who Conquers Self Succeeds"

Mark Szkolnicki

  • Sr. Member
  • ****
  • Posts: 433
  • Karma: +18/-0
    • View Profile
Re: Updated Uniaud32 build
« Reply #869 on: September 21, 2022, 07:56:38 pm »
Hi Paul!

Hi Mark,
Thanks - tonight, I'll aim to produce several builds, slowly replacing the 3.8.x pci/hda/* code with the 3.9.x pci/hda/* code - ideally one source file at a time.

Note: these are specifically directed at trying to understand what broken with his hardware between the sound code in linux 3.8, and linux 3.9.

Code: [Select]
hda_intel.c https://smedley.id.au/tmp/uniaud32-linux-3.9.11-20220921.zip

hda_auto_parser.c https://smedley.id.au/tmp/uniaud32-linux-3.9.11-20220921b.zip
hda_auto_parser.h
hda_codec.c
hda_codec.h
hda_local.h
hda_proc.c
hda_generic.c
hda_generic.h
patch_realtek.c

hda_beep.c https://smedley.id.au/tmp/uniaud32-linux-3.9.11-20220921c.zip
hda_beep.h


hda_eld.c https://smedley.id.au/tmp/uniaud32-linux-3.9.11-20220921d.zip

hda_hwdep.c https://smedley.id.au/tmp/uniaud32-linux-3.9.11-20220921e.zip

hda_jack.c https://smedley.id.au/tmp/uniaud32-linux-3.9.11-20220921f.zip
hda_jack.h



hda_intel_trace.h https://smedley.id.au/tmp/uniaud32-linux-3.9.11-20220921g.zip
hda_patch.h
hda_trace.h

If I had to bet, uniaud32-linux-3.9.11-20220921b.zip is the one that will break things. It's both the most obvious source files impacted, but also it's the largest code change - unfortunately there are interdependencies between those source files, so changing one at a time wasn't possible.

Hopefully I'm wrong, and it will be once of the smaller code changes.

Cheers,

Paul

Unfortunately Paul, your multichoice answer on this test was again correct (your on a roll, just in the hardest way  :P).

Things broke @ hda_auto_parser.c (your build uniaud-linux-3.9.11-2022-09-21b)

I did run ALL 7 builds, in sequence, and no sound was generated at all after I ran "b" - "a" (or the first file as you didn't label it as such) was fine.

To be thorough (just the troubleshooter in me  8) ), I ran all 7 both as a warm boot and a cold boot, just to make sure, like the Recon3D, that wasn't causing silences - same situation with all builds, "b" and beyond.

Like when I was helping Wim on the other project I've been actively involved in, I assume a few things, operating independently - obvious once one source code change is found that causes the silences, the subsequent source code changes would not work, but I tested all of them just in case.

I've generated testlogs for all, on the off chance more than one may tell you something (attached below). They are marked SCa to SCg

Best!

M
Vincit Que Se Vincit - "He Who Conquers Self Succeeds"