[alsa-devel] Incorrect model selection for HDA Intel based card.
Just had a user in #alsa on freenode, with an Asus P5GD1 mainboard. They couldnt seem to get surround sound working at all. It has an ALC880 based HDA-Intel sound device on board.
Now, the problem lies in the Susbsystem ID's (1043:814e). These match in patch_realtek.c as follows:
SND_PCI_QUIRK(0x1043, 0x814e, "ASUS", ALC880_ASUS)
The users mainboard (The ASUS P5GD1) actually has 6 output jacks, aswell as SPDIF, not the standard 3-jacks that the "ASUS" model quirk implies.
I talked this user through forcing the driver to use '6stack-digout' for his sound device, and things all work fine now.
Below is a link to his alsa-info output, with the problem fixed. (so you can check IDs etc)
Up until this point, i hadnt come across a wrong detection like that. Is there anyway to differentiate this users card from the normal "3stack" model that actually requires the "ASUS" model quirk ?
Thanks, Travis Place
At Mon, 19 May 2008 04:19:14 +1000, Travis Place wrote:
Just had a user in #alsa on freenode, with an Asus P5GD1 mainboard. They couldnt seem to get surround sound working at all. It has an ALC880 based HDA-Intel sound device on board.
Now, the problem lies in the Susbsystem ID's (1043:814e). These match in patch_realtek.c as follows:
SND_PCI_QUIRK(0x1043, 0x814e, "ASUS", ALC880_ASUS)
The users mainboard (The ASUS P5GD1) actually has 6 output jacks, aswell as SPDIF, not the standard 3-jacks that the "ASUS" model quirk implies.
I talked this user through forcing the driver to use '6stack-digout' for his sound device, and things all work fine now.
Actually, this was changed originally from ALC880_5STACK for the exactly same mobo (changeset 4798). I guess 6stack-digout wasn't tested by that user.
So, if 6stack-digout is confirmed to match better than now, it's fairly safe to replace the line. Care to send a patch with a proper changelog?
thanks,
Takashi
On Mon, 19 May 2008 10:09:00 pm Takashi Iwai wrote:
At Mon, 19 May 2008 04:19:14 +1000,
Travis Place wrote:
Just had a user in #alsa on freenode, with an Asus P5GD1 mainboard. They couldnt seem to get surround sound working at all. It has an ALC880 based HDA-Intel sound device on board.
Now, the problem lies in the Susbsystem ID's (1043:814e). These match in patch_realtek.c as follows:
SND_PCI_QUIRK(0x1043, 0x814e, "ASUS", ALC880_ASUS)
The users mainboard (The ASUS P5GD1) actually has 6 output jacks, aswell as SPDIF, not the standard 3-jacks that the "ASUS" model quirk implies.
I talked this user through forcing the driver to use '6stack-digout' for his sound device, and things all work fine now.
Actually, this was changed originally from ALC880_5STACK for the exactly same mobo (changeset 4798). I guess 6stack-digout wasn't tested by that user.
So, if 6stack-digout is confirmed to match better than now, it's fairly safe to replace the line. Care to send a patch with a proper changelog?
thanks,
Takashi _______________________________________________ Alsa-devel mailing list Alsa-devel@alsa-project.org http://mailman.alsa-project.org/mailman/listinfo/alsa-devel
Going to get the user to run a few more tests, to confirm all is working as it should (seems to be, on first glance) before submitting a patch/changelog.
Will let you know when ready.
Travis Place
Just noticed something very interesting in HG checkout..
SND_PCI_QUIRK(0x1043, 0x8196, "ASUS P5GD1", ALC880_6ST)
yet, the user states he has a P5GD1 but had the IDs 0x1043, 0x814e
in any event, it seems that 0x1043, 0x814e required ALC880_6ST_DIG
not sure what is going on here.
Travis Place.
At Tue, 20 May 2008 01:12:37 +1000, Travis Place wrote:
Just noticed something very interesting in HG checkout..
SND_PCI_QUIRK(0x1043, 0x8196, "ASUS P5GD1", ALC880_6ST)
yet, the user states he has a P5GD1 but had the IDs 0x1043, 0x814e
in any event, it seems that 0x1043, 0x814e required ALC880_6ST_DIG
not sure what is going on here.
The difference is the presense of SPDIF, so it was likely tested only without SPDIF. Using 6ST_DIG is fine even on hardwares without SPDIF. It'd just create useless mixer elements and PCM streams.
thanks,
Takashi
On Mon, 19 May 2008 10:09:00 pm Takashi Iwai wrote:
At Mon, 19 May 2008 04:19:14 +1000,
Travis Place wrote:
Just had a user in #alsa on freenode, with an Asus P5GD1 mainboard. They couldnt seem to get surround sound working at all. It has an ALC880 based HDA-Intel sound device on board.
Now, the problem lies in the Susbsystem ID's (1043:814e). These match in patch_realtek.c as follows:
SND_PCI_QUIRK(0x1043, 0x814e, "ASUS", ALC880_ASUS)
The users mainboard (The ASUS P5GD1) actually has 6 output jacks, aswell as SPDIF, not the standard 3-jacks that the "ASUS" model quirk implies.
I talked this user through forcing the driver to use '6stack-digout' for his sound device, and things all work fine now.
Actually, this was changed originally from ALC880_5STACK for the exactly same mobo (changeset 4798). I guess 6stack-digout wasn't tested by that user.
So, if 6stack-digout is confirmed to match better than now, it's fairly safe to replace the line. Care to send a patch with a proper changelog?
thanks,
Takashi _______________________________________________ Alsa-devel mailing list Alsa-devel@alsa-project.org http://mailman.alsa-project.org/mailman/listinfo/alsa-devel
As previously mentioned, it appears the HG code has an entry for the "ASUS P5GD1" mainboard. But its onboard sound vendor/device Ids dont match those of the person i was helping. I have double checked the pastebin post, and confirmed the mainboard model with the user though.
The _original_ alsa-info.sh run (before the model was forced to 6stack-digout) is as follows..
There it shows the vendor/device ids as 0x1043, 0x814e . So according to the current HG code, those IDs would select the "asus" model. Also, in the same, current HG code, we have..
SND_PCI_QUIRK(0x1043, 0x8196, "ASUS P5GD1", ALC880_6ST)
This shows the correct model name for the mainboard, but (from my experience) wrong vendor/device ids.
Perhaps we have to check with the person who originally added that P5GD1 quirk ?
I have made up a patch, but want to confirm all this information before proceeding. Please let me know your thoughts.
Travis Place
At Tue, 20 May 2008 01:43:31 +1000, Travis Place wrote:
On Mon, 19 May 2008 10:09:00 pm Takashi Iwai wrote:
At Mon, 19 May 2008 04:19:14 +1000,
Travis Place wrote:
Just had a user in #alsa on freenode, with an Asus P5GD1 mainboard. They couldnt seem to get surround sound working at all. It has an ALC880 based HDA-Intel sound device on board.
Now, the problem lies in the Susbsystem ID's (1043:814e). These match in patch_realtek.c as follows:
SND_PCI_QUIRK(0x1043, 0x814e, "ASUS", ALC880_ASUS)
The users mainboard (The ASUS P5GD1) actually has 6 output jacks, aswell as SPDIF, not the standard 3-jacks that the "ASUS" model quirk implies.
I talked this user through forcing the driver to use '6stack-digout' for his sound device, and things all work fine now.
Actually, this was changed originally from ALC880_5STACK for the exactly same mobo (changeset 4798). I guess 6stack-digout wasn't tested by that user.
So, if 6stack-digout is confirmed to match better than now, it's fairly safe to replace the line. Care to send a patch with a proper changelog?
thanks,
Takashi _______________________________________________ Alsa-devel mailing list Alsa-devel@alsa-project.org http://mailman.alsa-project.org/mailman/listinfo/alsa-devel
As previously mentioned, it appears the HG code has an entry for the "ASUS P5GD1" mainboard. But its onboard sound vendor/device Ids dont match those of the person i was helping. I have double checked the pastebin post, and confirmed the mainboard model with the user though.
The _original_ alsa-info.sh run (before the model was forced to 6stack-digout) is as follows..
There it shows the vendor/device ids as 0x1043, 0x814e .
1043:814e is also P5GD1. http://www.linuxquestions.org/questions/showthread.php?p=2556497#post2556497
It was changed to ALC880_ASUS based on the report in the above.
So according to the current HG code, those IDs would select the "asus" model. Also, in the same, current HG code, we have..
SND_PCI_QUIRK(0x1043, 0x8196, "ASUS P5GD1", ALC880_6ST)
This shows the correct model name for the mainboard, but (from my experience) wrong vendor/device ids.
Perhaps we have to check with the person who originally added that P5GD1 quirk ?
I have made up a patch, but want to confirm all this information before proceeding. Please let me know your thoughts.
I believe it's fine to change to 6stack model if the tester's board is also P5GD1 indeed. Of course, it's possible that ASUS may change the hardware design completely for the same model number with PCI SSID, but I hope they aren't so crazy :)
thanks,
Takashi
On Tue, 20 May 2008 1:54:42 am Takashi Iwai wrote:
At Tue, 20 May 2008 01:43:31 +1000,
Travis Place wrote:
On Mon, 19 May 2008 10:09:00 pm Takashi Iwai wrote:
At Mon, 19 May 2008 04:19:14 +1000,
Travis Place wrote:
Just had a user in #alsa on freenode, with an Asus P5GD1 mainboard. They couldnt seem to get surround sound working at all. It has an ALC880 based HDA-Intel sound device on board.
Now, the problem lies in the Susbsystem ID's (1043:814e). These match in patch_realtek.c as follows:
SND_PCI_QUIRK(0x1043, 0x814e, "ASUS", ALC880_ASUS)
The users mainboard (The ASUS P5GD1) actually has 6 output jacks, aswell as SPDIF, not the standard 3-jacks that the "ASUS" model quirk implies.
I talked this user through forcing the driver to use '6stack-digout' for his sound device, and things all work fine now.
Actually, this was changed originally from ALC880_5STACK for the exactly same mobo (changeset 4798). I guess 6stack-digout wasn't tested by that user.
So, if 6stack-digout is confirmed to match better than now, it's fairly safe to replace the line. Care to send a patch with a proper changelog?
thanks,
Takashi _______________________________________________ Alsa-devel mailing list Alsa-devel@alsa-project.org http://mailman.alsa-project.org/mailman/listinfo/alsa-devel
As previously mentioned, it appears the HG code has an entry for the "ASUS P5GD1" mainboard. But its onboard sound vendor/device Ids dont match those of the person i was helping. I have double checked the pastebin post, and confirmed the mainboard model with the user though.
The _original_ alsa-info.sh run (before the model was forced to 6stack-digout) is as follows..
There it shows the vendor/device ids as 0x1043, 0x814e .
1043:814e is also P5GD1. http://www.linuxquestions.org/questions/showthread.php?p=2556497#post25564 97
It was changed to ALC880_ASUS based on the report in the above.
So according to the current HG code, those IDs would select the "asus" model. Also, in the same, current HG code, we have..
SND_PCI_QUIRK(0x1043, 0x8196, "ASUS P5GD1", ALC880_6ST)
This shows the correct model name for the mainboard, but (from my experience) wrong vendor/device ids.
Perhaps we have to check with the person who originally added that P5GD1 quirk ?
I have made up a patch, but want to confirm all this information before proceeding. Please let me know your thoughts.
I believe it's fine to change to 6stack model if the tester's board is also P5GD1 indeed. Of course, it's possible that ASUS may change the hardware design completely for the same model number with PCI SSID, but I hope they aren't so crazy :)
thanks,
Takashi _______________________________________________ Alsa-devel mailing list Alsa-devel@alsa-project.org http://mailman.alsa-project.org/mailman/listinfo/alsa-devel
Below (as well as attached) is my patch, for the 0x814e variant.
diff -r e2b60934d1c8 pci/hda/patch_realtek.c --- a/pci/hda/patch_realtek.c Mon May 19 12:32:25 2008 +0200 +++ b/pci/hda/patch_realtek.c Tue May 20 01:50:04 2008 +1000 @@ -2981,7 +2981,7 @@ static struct snd_pci_quirk alc880_cfg_t /* SND_PCI_QUIRK(0x1043, 0x1964, "ASUS", ALC880_ASUS_DIG), */ SND_PCI_QUIRK(0x1043, 0x1973, "ASUS", ALC880_ASUS_DIG), SND_PCI_QUIRK(0x1043, 0x19b3, "ASUS", ALC880_ASUS_DIG), - SND_PCI_QUIRK(0x1043, 0x814e, "ASUS", ALC880_ASUS), + SND_PCI_QUIRK(0x1043, 0x814e, "ASUS P5GD1 w/SPDIF", ALC880_6ST_DIG), SND_PCI_QUIRK(0x1043, 0x8181, "ASUS P4GPL", ALC880_ASUS_DIG), SND_PCI_QUIRK(0x1043, 0x8196, "ASUS P5GD1", ALC880_6ST), SND_PCI_QUIRK(0x1043, 0x81b4, "ASUS", ALC880_6ST),
Changelog should be something like:
Corrected the model assignment for the ASUS P5GD1 w/SPDIF after reports of surround sound not being possible.
Let me know if you need a more detailed changelog entry.
Travis Place
At Tue, 20 May 2008 02:01:02 +1000, Travis Place wrote:
On Tue, 20 May 2008 1:54:42 am Takashi Iwai wrote:
At Tue, 20 May 2008 01:43:31 +1000,
Travis Place wrote:
On Mon, 19 May 2008 10:09:00 pm Takashi Iwai wrote:
At Mon, 19 May 2008 04:19:14 +1000,
Travis Place wrote:
Just had a user in #alsa on freenode, with an Asus P5GD1 mainboard. They couldnt seem to get surround sound working at all. It has an ALC880 based HDA-Intel sound device on board.
Now, the problem lies in the Susbsystem ID's (1043:814e). These match in patch_realtek.c as follows:
SND_PCI_QUIRK(0x1043, 0x814e, "ASUS", ALC880_ASUS)
The users mainboard (The ASUS P5GD1) actually has 6 output jacks, aswell as SPDIF, not the standard 3-jacks that the "ASUS" model quirk implies.
I talked this user through forcing the driver to use '6stack-digout' for his sound device, and things all work fine now.
Actually, this was changed originally from ALC880_5STACK for the exactly same mobo (changeset 4798). I guess 6stack-digout wasn't tested by that user.
So, if 6stack-digout is confirmed to match better than now, it's fairly safe to replace the line. Care to send a patch with a proper changelog?
thanks,
Takashi _______________________________________________ Alsa-devel mailing list Alsa-devel@alsa-project.org http://mailman.alsa-project.org/mailman/listinfo/alsa-devel
As previously mentioned, it appears the HG code has an entry for the "ASUS P5GD1" mainboard. But its onboard sound vendor/device Ids dont match those of the person i was helping. I have double checked the pastebin post, and confirmed the mainboard model with the user though.
The _original_ alsa-info.sh run (before the model was forced to 6stack-digout) is as follows..
There it shows the vendor/device ids as 0x1043, 0x814e .
1043:814e is also P5GD1. http://www.linuxquestions.org/questions/showthread.php?p=2556497#post25564 97
It was changed to ALC880_ASUS based on the report in the above.
So according to the current HG code, those IDs would select the "asus" model. Also, in the same, current HG code, we have..
SND_PCI_QUIRK(0x1043, 0x8196, "ASUS P5GD1", ALC880_6ST)
This shows the correct model name for the mainboard, but (from my experience) wrong vendor/device ids.
Perhaps we have to check with the person who originally added that P5GD1 quirk ?
I have made up a patch, but want to confirm all this information before proceeding. Please let me know your thoughts.
I believe it's fine to change to 6stack model if the tester's board is also P5GD1 indeed. Of course, it's possible that ASUS may change the hardware design completely for the same model number with PCI SSID, but I hope they aren't so crazy :)
thanks,
Takashi _______________________________________________ Alsa-devel mailing list Alsa-devel@alsa-project.org http://mailman.alsa-project.org/mailman/listinfo/alsa-devel
Below (as well as attached) is my patch, for the 0x814e variant.
diff -r e2b60934d1c8 pci/hda/patch_realtek.c --- a/pci/hda/patch_realtek.c Mon May 19 12:32:25 2008 +0200 +++ b/pci/hda/patch_realtek.c Tue May 20 01:50:04 2008 +1000 @@ -2981,7 +2981,7 @@ static struct snd_pci_quirk alc880_cfg_t /* SND_PCI_QUIRK(0x1043, 0x1964, "ASUS", ALC880_ASUS_DIG), */ SND_PCI_QUIRK(0x1043, 0x1973, "ASUS", ALC880_ASUS_DIG), SND_PCI_QUIRK(0x1043, 0x19b3, "ASUS", ALC880_ASUS_DIG),
- SND_PCI_QUIRK(0x1043, 0x814e, "ASUS", ALC880_ASUS),
- SND_PCI_QUIRK(0x1043, 0x814e, "ASUS P5GD1 w/SPDIF", ALC880_6ST_DIG), SND_PCI_QUIRK(0x1043, 0x8181, "ASUS P4GPL", ALC880_ASUS_DIG), SND_PCI_QUIRK(0x1043, 0x8196, "ASUS P5GD1", ALC880_6ST), SND_PCI_QUIRK(0x1043, 0x81b4, "ASUS", ALC880_6ST),
Changelog should be something like:
Corrected the model assignment for the ASUS P5GD1 w/SPDIF after reports of surround sound not being possible.
Let me know if you need a more detailed changelog entry.
That's enough for this simple change, but it'd be appreciated if you can give your sign-off.
thanks,
Takashi
participants (2)
-
Takashi Iwai -
Travis Place