After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 761031 - Newer audio device for modern windows
Newer audio device for modern windows
Status: RESOLVED OBSOLETE
Product: gnome-boxes
Classification: Applications
Component: general
3.18.x
Other Linux
: Normal enhancement
: --
Assigned To: GNOME Boxes maintainer(s)
GNOME Boxes maintainer(s)
Depends on:
Blocks:
 
 
Reported: 2016-01-24 00:16 UTC by Daniel Aleksandersen
Modified: 2018-01-11 10:31 UTC
See Also:
GNOME target: ---
GNOME version: ---



Description Daniel Aleksandersen 2016-01-24 00:16:53 UTC
There is no 64-bit driver for AC97 audio cards, and drivers are not available for Windows 8 nor 10.

“qemu -device intel-hda“ seems to work for everyone on all OSes, as per experiences shared on around the web. (Aka. my Google searches were everyone’s problems seem to be fixed by “changing to Intel HDA” in quem and VirtualBox.)

The audio device is non-configurable per the below commit. If you don’t want to change the default audio device, please make it configurable to support newer operating systems..
https://mail.gnome.org/archives/commits-list/2011-December/msg08329.html

VirtualBox offers these choices: intel-hda, ac97, and soundblaster16. VirtualBox defaults to intel-hda for Windows systems and ac97 for Linux systems.
Comment 1 Zeeshan Ali 2016-01-24 16:43:54 UTC
(In reply to Daniel Aleksandersen from comment #0)
> There is no 64-bit driver for AC97 audio cards, and drivers are not
> available for Windows 8 nor 10.
> 
> “qemu -device intel-hda“ seems to work for everyone on all OSes, as per
> experiences shared on around the web. (Aka. my Google searches were
> everyone’s problems seem to be fixed by “changing to Intel HDA” in quem and
> VirtualBox.)
> 
> The audio device is non-configurable per the below commit. If you don’t want
> to change the default audio device, please make it configurable to support
> newer operating systems..
> https://mail.gnome.org/archives/commits-list/2011-December/msg08329.html

Configuration option is typically a work-around and is very much so in this case. The correct fix would be to change the audio device for these OSes in libosinfo database.
Comment 2 Daniel Aleksandersen 2016-02-18 12:50:15 UTC
I wrote up how to modify your Windows box installation to get working audio, for any users who’re waiting for a fix to this issue.
https://www.slightfuture.com/how-to/win10-in-gnome-boxes
Comment 3 Felipe Borges 2017-04-11 16:02:20 UTC
I just installed a Windows 10 box specifically to reproduce your issue.

It sounds good ;).I could hear Cortana speaking in the bootstrapping dialog and after as well. Everything works fine out of the box.

I see in the machine xml that it picked up the right sound model option:
    <sound model='ich6'>
      <address type='pci' domain='0x0000' bus='0x00' slot='0x04' function='0x0'/>
    </sound>

Are you still experiencing the issue?

I am testing it with:
gnome-boxes-3.24.0 in Fedora 26
Windows 10 Home x64
Comment 4 Édouard WILLISSECK 2017-04-11 17:30:21 UTC
I set up a Windows 7 VM on Debian Testing.

It used 'ac97' and the audio was not working. I had to manually edit the xml file to change it to 'ich6'.

As I said I used Debian Testing and it was gnome-boxes 3.22.4

I guess it is fixed in the new version. Just need to wait for Debian to update.

(In reply to Felipe Borges from comment #3)
> I just installed a Windows 10 box specifically to reproduce your issue.
> 
> It sounds good ;).I could hear Cortana speaking in the bootstrapping dialog
> and after as well. Everything works fine out of the box.
> 
> I see in the machine xml that it picked up the right sound model option:
>     <sound model='ich6'>
>       <address type='pci' domain='0x0000' bus='0x00' slot='0x04'
> function='0x0'/>
>     </sound>
> 
> Are you still experiencing the issue?
> 
> I am testing it with:
> gnome-boxes-3.24.0 in Fedora 26
> Windows 10 Home x64
Comment 5 Felipe Borges 2017-04-12 14:31:58 UTC
I tested this time with an Windows 7 box and I can reproduce your issue.

I see in the osinfo-db/data/os/microsoft.com/win-7.xml.in that the correct device (ich6) is defined, but it is not getting picked up when installing the win-7 guest. Updating the sound entry in the xml with virsh edit fixed it for me as well.

I reported this bug against libosinfo/osinfo-db: https://bugzilla.redhat.com/show_bug.cgi?id=1441719
Comment 6 Chris Rainey 2017-12-22 19:36:55 UTC
Problem reproduced:

Windows 10 Home(x86-64)
Ver. 1709
Build 16299.125

Changed "ac97" ---> "ich6" in boxes-*.xml w/ 'virsh edit' command to fix.


Ubuntu 17.04

chris@CKR-1:~$ gnome-boxes --version
3.24.0
Comment 7 Felipe Borges 2018-01-08 13:52:11 UTC
(In reply to Chris Rainey from comment #6)
> Problem reproduced:
> 
> Windows 10 Home(x86-64)
> Ver. 1709
> Build 16299.125
> 
> Changed "ac97" ---> "ich6" in boxes-*.xml w/ 'virsh edit' command to fix.
> 
> 
> Ubuntu 17.04
> 
> chris@CKR-1:~$ gnome-boxes --version
> 3.24.0

The osinfo data seems to be correctly pointing to the ich6 driver but we fallback to ac97 whenever the Operating System is not identified.

The reasoning is discussed in Bug 672872. I'd have to conduct some research to conclude whether the assumptions posted in this bug are still valid.

Therefore, would you mind sharing the output of $ osinfo-detect your_windows_image_file.iso ?
Comment 8 Chris Rainey 2018-01-08 18:38:46 UTC
(In reply to Felipe Borges from comment #7)
> (In reply to Chris Rainey from comment #6)
> > Problem reproduced:
> > 
> > Windows 10 Home(x86-64)
> > Ver. 1709
> > Build 16299.125
> > 
> > Changed "ac97" ---> "ich6" in boxes-*.xml w/ 'virsh edit' command to fix.
> > 
> > 
> > Ubuntu 17.04
> > 
> > chris@CKR-1:~$ gnome-boxes --version
> > 3.24.0
> 
> The osinfo data seems to be correctly pointing to the ich6 driver but we
> fallback to ac97 whenever the Operating System is not identified.
> 
> The reasoning is discussed in Bug 672872. I'd have to conduct some research
> to conclude whether the assumptions posted in this bug are still valid.
> 
> Therefore, would you mind sharing the output of $ osinfo-detect
> your_windows_image_file.iso ?


chris@CKR-1:~$ osinfo-detect Archives/ISOS/Win10_1709_English_x64.iso 
Media is bootable.
chris@CKR-1:~$ 


Thanks for looking into this!
Comment 9 Felipe Borges 2018-01-09 14:10:50 UTC
(In reply to Chris Rainey from comment #8)
> (In reply to Felipe Borges from comment #7)
> > Therefore, would you mind sharing the output of $ osinfo-detect
> > your_windows_image_file.iso ?
> 
> 
> chris@CKR-1:~$ osinfo-detect Archives/ISOS/Win10_1709_English_x64.iso 
> Media is bootable.

That's it. Your image file is not detected by libosinfo, causing Boxes to fallback to the AC97 driver which we assign to unknown (undetected) operating systems. See https://git.gnome.org/browse/gnome-boxes/tree/src/vm-configurator.vala#n357

After discussing with @teuf on IRC, it seems that the right approach is to switch the fallback to ich6 and patch older Windows versions in libosinfo to explicitly pick the AC97 driver.

I will patch Boxes as soon as we get the libosinfo patches.

Expect this to be resolved/fixed for our next stable release: 3.28.
Comment 10 Christophe Fergeau 2018-01-09 14:22:26 UTC
(In reply to Chris Rainey from comment #8)
> 
> chris@CKR-1:~$ osinfo-detect Archives/ISOS/Win10_1709_English_x64.iso 
> Media is bootable.

Could you provide the output of
$ isoinfo -d -i Archives/ISOS/Win10_1709_English_x64.iso
?

isoinfo is part of the genisoimage/cdrkit package.
Comment 11 Chris Rainey 2018-01-09 15:00:52 UTC
(In reply to Christophe Fergeau from comment #10)
> (In reply to Chris Rainey from comment #8)
> > 
> > chris@CKR-1:~$ osinfo-detect Archives/ISOS/Win10_1709_English_x64.iso 
> > Media is bootable.
> 
> Could you provide the output of
> $ isoinfo -d -i Archives/ISOS/Win10_1709_English_x64.iso
> ?
> 
> isoinfo is part of the genisoimage/cdrkit package.


You, Sir --- are a God among Devs!


chris@CKR-1:~$ isoinfo -d -i Archives/ISOS/Win10_1709_English_x64.iso
CD-ROM is in ISO 9660 format
System id: 
Volume id: CCCOMA_X64FRE_EN-US_DV9
Volume set id: CCCOMA_X64FRE_EN-US_DV9
Publisher id: MICROSOFT CORPORATION
Data preparer id: MICROSOFT CORPORATION, ONE MICROSOFT WAY, REDMOND WA 98052, (425) 882-8080
Application id: CDIMAGE 2.56 (01/01/2005 TM)
Copyright File id: 
Abstract File id: 
Bibliographic File id: 
Volume set size is: 1
Volume set sequence number is: 1
Logical block size is: 2048
Volume size is: 2293634
El Torito VD version 1 found, boot catalog is in sector 22
NO Joliet present
NO Rock Ridge present
Eltorito validation header:
    Hid 1
    Arch 0 (x86)
    ID 'Microsoft Corporation'
    Key 55 AA
    Eltorito defaultboot header:
        Bootid 88 (bootable)
        Boot media 0 (No Emulation Boot)
        Load segment 0
        Sys type 0
        Nsect 8
        Bootoff 20C 524
chris@CKR-1:~$
Comment 12 Christophe Fergeau 2018-01-09 16:09:45 UTC
(In reply to Chris Rainey from comment #11)
> (In reply to Christophe Fergeau from comment #10)
> > (In reply to Chris Rainey from comment #8)
> > > 
> > > chris@CKR-1:~$ osinfo-detect Archives/ISOS/Win10_1709_English_x64.iso 
> > > Media is bootable.
> > 
> > Could you provide the output of
> > $ isoinfo -d -i Archives/ISOS/Win10_1709_English_x64.iso
> > ?
> > 
> > isoinfo is part of the genisoimage/cdrkit package.
> 
> 
> You, Sir --- are a God among Devs!
> 
> 
> chris@CKR-1:~$ isoinfo -d -i Archives/ISOS/Win10_1709_English_x64.iso

Thanks! osinfodb from git seems to already correctly identify this image :)
Comment 13 GNOME Infrastructure Team 2018-01-11 10:31:15 UTC
-- GitLab Migration Automatic Message --

This bug has been migrated to GNOME's GitLab instance and has been closed from further activity.

You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.gnome.org/GNOME/gnome-boxes/issues/78.