Author Topic: Updated Uniaud32 build  (Read 407186 times)

Mark Szkolnicki

  • Sr. Member
  • ****
  • Posts: 433
  • Karma: +18/-0
    • View Profile
Re: Updated Uniaud32 build
« Reply #780 on: September 08, 2022, 08:04:57 pm »
Hey Paul!

Thanks for testing Mike, now that most of the uniaud32 issues appear to be resolved, I'll get back to stabilising smbd - in particular when sharing resources that are Netdrive mounts.

Tested UniAud32 5.15.65 20220908 on both systems:

Manticore II - AD1988B - still functioning perfectly - system sounds present and PM123 works perfectly

Hyperfire - still no sound with the ALC892 chipset, neither system sounds nor PM123  - tested with both the Release and Debug drivers. UniMix still reports 10 volume sections - volumes were set to max.

I've attached an updated Testlog for Hyperfire, which was generated today with 0908 and the Debug version you built. If you wish I will generate a TRAC ticket.

Best!

M
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 #781 on: September 09, 2022, 12:39:32 am »
Hey Mark,

I've attached an updated Testlog for Hyperfire, which was generated today with 0908 and the Debug version you built. If you wish I will generate a TRAC ticket.

I just had a quick look at the log, the part that seems strange to me is:
Code: [Select]
codec #1 probed OK

snd_hda_codec_device_init: entry

snd_hda_codec_device_new: entry

Unable to bind the codec

Cannot probe codecs, giving up

I need to check against another system, but I don't believe these lines are normal. I'll do some checks tonight and also chekc the relevant code to see if here is anything enlightening.

Cheers,

Paul

Mark Szkolnicki

  • Sr. Member
  • ****
  • Posts: 433
  • Karma: +18/-0
    • View Profile
Re: Updated Uniaud32 build
« Reply #782 on: September 09, 2022, 03:15:07 pm »
Hi Paul

Hey Mark,

I've attached an updated Testlog for Hyperfire, which was generated today with 0908 and the Debug version you built. If you wish I will generate a TRAC ticket.

I just had a quick look at the log, the part that seems strange to me is:
Code: [Select]
codec #1 probed OK

snd_hda_codec_device_init: entry

snd_hda_codec_device_new: entry

Unable to bind the codec

Cannot probe codecs, giving up

I need to check against another system, but I don't believe these lines are normal. I'll do some checks tonight and also chekc the relevant code to see if here is anything enlightening.

Cheers,

Paul

I generated a testlog for UniAud on the working system (Manticore II) which I've attached below, if that helps out at all - not the same chipset, but may help for comparison purposes.

M
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 #783 on: September 09, 2022, 11:29:52 pm »
I generated a testlog for UniAud on the working system (Manticore II) which I've attached below, if that helps out at all - not the same chipset, but may help for comparison purposes.

It saved me generating one myself :)

As I suspected, your other log has:
Code: [Select]
codec #0 probed OK

codec #3 probed OK

snd_hda_codec_device_init: entry

snd_hda_codec_device_new: entry

snd_hda_codec_device_init: entry

snd_hda_codec_device_new: entry

Unfortunately, I have no real ideas right now on why this is happening :(

Paul Smedley

  • Hero Member
  • *****
  • Posts: 2095
  • Karma: +160/-0
    • View Profile
Re: Updated Uniaud32 build
« Reply #784 on: September 10, 2022, 12:16:11 am »
@Mark - I don't really expect this to help - but this does update some functions that are mentioned in https://lkml.org/lkml/2021/11/15/2038

https://smedley.id.au/tmp/uniaud32-linux-5.15.65-20220910.zip

I also added a #define for a HDMI codec - which who knows, might help HDMI audio - I don't have an easy way to test this right now.
« Last Edit: September 10, 2022, 02:28:07 am by Paul Smedley »

Mark Szkolnicki

  • Sr. Member
  • ****
  • Posts: 433
  • Karma: +18/-0
    • View Profile
Re: Updated Uniaud32 build
« Reply #785 on: September 10, 2022, 02:06:46 am »
Hey Paul

@Mark - I don't really expect this to help - but this does update some functions that are mentioned in https://lkml.org/lkml/2021/11/15/2038

https://smedley.id.au/tmp/uniaud32-linux-5.15.65-20220910.zip

I doubt you claim omnipotence, and I certainly don't, Sir - but I'll try the latest update.

Like always, at least we know what doesn't work ............

In emergency's we always try bright ideas when common sense fails - sometimes the most absurd succeeds ......... why is a mystery to me, even after 42 years doing this ......... !

M
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 #786 on: September 10, 2022, 02:24:57 am »
I also just refreshed the zip to try and load the azx driver (aka hda) driver first - previously it would iterate through ~10 other drivers first - given the prevalence of hda hardware these days, I figured it made sense to try HDA first - might shave a few milliseconds off of the boot time on HDA hardware :P

Paul Smedley

  • Hero Member
  • *****
  • Posts: 2095
  • Karma: +160/-0
    • View Profile
Re: Updated Uniaud32 build
« Reply #787 on: September 10, 2022, 02:26:25 am »
Hey Mark!

@Mark - I don't really expect this to help - but this does update some functions that are mentioned in https://lkml.org/lkml/2021/11/15/2038

https://smedley.id.au/tmp/uniaud32-linux-5.15.65-20220910.zip

I doubt you claim omnipotence, and I certainly don't, Sir - but I'll try the latest update.

Like always, at least we know what doesn't work ............

In emergency's we always try bright ideas when common sense fails - sometimes the most absurd succeeds ......... why is a mystery to me, even after 42 years doing this ......... !

I certainly don't claim omnipotence :) I'm just a hacker who enjoys trying to make things work, even though it consumes time I could probably better spend elsewhere :P

edit: https://smedley.id.au/tmp/uniaud32-linux-5.15.65-20220910b.zip implements pci_intx - which was previously stubbed. This is only used for HDA hardware - hda continues to work for me in Virtualbox and on a T470s (ie bare metal).

Cheers,

Paul
« Last Edit: September 10, 2022, 02:52:15 am by Paul Smedley »

Martin Iturbide

  • OS2World NewsMaster
  • Global Moderator
  • Hero Member
  • *****
  • Posts: 4754
  • Karma: +41/-1
  • Your Friend Wil Declares...
    • View Profile
    • Martin's Personal Blog
Re: Updated Uniaud32 build
« Reply #788 on: September 10, 2022, 01:36:15 pm »
edit: https://smedley.id.au/tmp/uniaud32-linux-5.15.65-20220910b.zip implements pci_intx - which was previously stubbed. This is only used for HDA hardware - hda continues to work for me in Virtualbox and on a T470s (ie bare metal).

Hi Paul
The link is not working for me.
Regards
Martin Iturbide
OS2World NewsMaster
... just share the dream.

Mark Szkolnicki

  • Sr. Member
  • ****
  • Posts: 433
  • Karma: +18/-0
    • View Profile
Re: Updated Uniaud32 build
« Reply #789 on: September 10, 2022, 04:33:08 pm »
Hey Paul!

Hey Mark!

@Mark - I don't really expect this to help - but this does update some functions that are mentioned in https://lkml.org/lkml/2021/11/15/2038

https://smedley.id.au/tmp/uniaud32-linux-5.15.65-20220910.zip

I doubt you claim omnipotence, and I certainly don't, Sir - but I'll try the latest update.

Like always, at least we know what doesn't work ............

In emergency's we always try bright ideas when common sense fails - sometimes the most absurd succeeds ......... why is a mystery to me, even after 42 years doing this ......... !

I certainly don't claim omnipotence :) I'm just a hacker who enjoys trying to make things work, even though it consumes time I could probably better spend elsewhere :P

edit: https://smedley.id.au/tmp/uniaud32-linux-5.15.65-20220910b.zip implements pci_intx - which was previously stubbed. This is only used for HDA hardware - hda continues to work for me in Virtualbox and on a T470s (ie bare metal).

Cheers,

Paul

Got tied up last night.

Ahhhh, a man after my own heart ............

I got into first the crazy business of Hazmat (Hazardous Materials) response, then emergency response in general because I'm a troubleshooter to the core - I want things to work, and in emergencies, you first have to stabilize the situation, and then scratch your head as to how to clean-up sometimes very unusual messes and troubles .........

It takes patient and a very thick skin in some situations ........... but after four decades I still love the craziness ............ !

As I've said many times, I'm not a programmer (but thinking of learning REXX this fall, to see if I can assist in pushing another project forward). But I get wanting to drop more mundane things to take on a challenge. When things start working its always a rush, having accomplished something that potentially no one else has done

Whether we can get ALC892 to work or not is irrelevant - but I appreciate you scratching your head and trying to figure it out ...........

And as Martin mentioned, the 10b link is coming up as a 404 - perhaps you took it down, after having yet another bright crazy idea .........

In the meantime will try the 10 release anyway and report back sometime today.

BTW Anne (my wife) and I were out visiting her cousin, who is a dog breeder, and were absolutely mobbed by a herd of 5 Miniature Schnauzers on Thursday, who took an instant liking to us - nearly took two of those middle weight mustached cuties home - which would probably have upset the current feline household of two .......... !

Best!

M


« Last Edit: September 10, 2022, 06:11:41 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 #790 on: September 10, 2022, 07:06:54 pm »
Heh Paul!

@Mark - I don't really expect this to help - but this does update some functions that are mentioned in https://lkml.org/lkml/2021/11/15/2038

https://smedley.id.au/tmp/uniaud32-linux-5.15.65-20220910.zip

I also added a #define for a HDMI codec - which who knows, might help HDMI audio - I don't have an easy way to test this right now.

Tried 5.15.65 20220910 on Hyperfire with the same results - didn't break anything but still no sound.

The IRQ 19 and 16 entries you pointed out are still in the latest log I generated today where "giving up" was mentioned regarding the probe

Skimming down the testlog I also noted these entries - don't know if they have any bearing on the problem but thought I'd mention them: 

realtek: No valid SSID, checking pincfg 0x4005e601 for NID 0x1d

realtek: Enabling init ASM_ID=0xe601 CODEC_ID=10ec0892

The problem may be related somehow specifically to the Realtek HDA codec (not being a programmer I'm not sure If I'm using the correct terminology here), and maybe to the SSID for the ALC892 chipset not being recognized by same. Am I making any sense?

M


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 #791 on: September 10, 2022, 11:44:42 pm »
edit: https://smedley.id.au/tmp/uniaud32-linux-5.15.65-20220910b.zip implements pci_intx - which was previously stubbed. This is only used for HDA hardware - hda continues to work for me in Virtualbox and on a T470s (ie bare metal).

Hi Paul
The link is not working for me.
Regards

oops, fixed now.

Mark Szkolnicki

  • Sr. Member
  • ****
  • Posts: 433
  • Karma: +18/-0
    • View Profile
Re: Updated Uniaud32 build
« Reply #792 on: September 11, 2022, 12:16:56 am »
Hey Paul!

edit: https://smedley.id.au/tmp/uniaud32-linux-5.15.65-20220910b.zip implements pci_intx - which was previously stubbed. This is only used for HDA hardware - hda continues to work for me in Virtualbox and on a T470s (ie bare metal).

Hi Paul
The link is not working for me.
Regards

oops, fixed now.

Tried 5.15.65 20220910b on Hyperfire and exactly the same as 10 in the previous builds. (no sound - 10 sound sections reported under PMUnimix)

Generated a Testlog and skimmed it - the sections we discussed previously are still exactly the same
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 #793 on: September 11, 2022, 02:16:33 am »
Hey Mark, I'm unfortunately pretty much out of ideas for Hyperfire. It would be interesting to boot that system from an Ubuntu 22.04.1 live image (https://releases.ubuntu.com/22.04/) - that release runs a 5.15 kernel so it should be pretty similar code.

Edit: I just noticed something interesting.... the output of unimix -card for Hyperfire is:
Card info for adapter 0:
  num: 0
  id: Generic
  driver: HDA-Intel
  name: HD-Audio Generic
  longname: HD-Audio Generic at 0xfc700000 irq 16
  mixer: Realtek ALC892
  componenets: HDA:10ec0892,1043841b,00100302

Other systems (such as Mantiscore and others) have:
Card info for adapter 0:
  num: 0
  id: NVidia
  driver: HDA-Intel
  name: HDA NVidia
  longname: HDA NVidia at 0xfe020000 irq 22
  mixer: Analog Devices AD1988B
  componenets: HDA:11d4198b,1043829b,00100400 HDA:10de0002,10de0101,00100000

I wonder if this is part of the problem?

Edit2: perhaps not... from the code:
   /* AMD/ATI Generic, PCI class code and Vendor ID for HD Audio */
   { PCI_DEVICE(PCI_VENDOR_ID_ATI, PCI_ANY_ID),
     .class = PCI_CLASS_MULTIMEDIA_HD_AUDIO << 8,
     .class_mask = 0xffffff,
     .driver_data = AZX_DRIVER_GENERIC | AZX_DCAPS_PRESET_ATI_HDMI },
   { PCI_DEVICE(PCI_VENDOR_ID_AMD, PCI_ANY_ID),
     .class = PCI_CLASS_MULTIMEDIA_HD_AUDIO << 8,
     .class_mask = 0xffffff,
     .driver_data = AZX_DRIVER_GENERIC | AZX_DCAPS_PRESET_ATI_HDMI },

So if the PCI Vendor ID matches AMD, it uses AZX_DRIVER_GENERIC - but it could still be something to do with the use of AZX_DRIVER_GENERIC vs the other specific drivers...
« Last Edit: September 11, 2022, 02:50:50 am by Paul Smedley »

Lars

  • Hero Member
  • *****
  • Posts: 1277
  • Karma: +65/-0
    • View Profile
Re: Updated Uniaud32 build
« Reply #794 on: September 11, 2022, 10:46:22 am »
I am not yet out of ideas:
if it is still the "giving up" error, I'd make routine "azx_probe" the same as "azx_probe_continue": call in both routines "pci_set_drvdata(PCI,NULL)" and "snd_card_free(card)" on the out_free error exit point. My suspicion is that the probing is aborted prematurely.
That's because azx_probe allocates a card but if it determines later on that there was an error, it does not reset the card info (the same explanation as for "azx_probe_continue" should be true).
I see another risk (maybe): "azx_,probe_continue" is called from within "azx_probe". But both call "snd_card_free" on error exit. Let's hope that is not a problem.
All that in file hda_intel.c.