[alsa-devel] [PATCH] add new Fermi 5xx codec IDs to snd-hda
Hello,
It's my first patch/post here, i'm not sure if this list is the right place to suggest this patch.
To follow up a previous commit on snd-hda : http://git.kernel.org/?p=linux/kernel/git/tiwai/sound-2.6.git;a=commitdiff;h...
I added a new entry in patch_hdmi.c to get sound working on my own 560 Ti card :
--- patch_hdmi.c 2011-02-15 22:29:51.000000000 +0100 +++ patch_hdmi_GTX-560Ti.c 2011-02-27 23:52:56.984613414 +0100 @@ -1634,6 +1634,7 @@ { .id = 0x10de0012, .name = "GPU 12 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, { .id = 0x10de0013, .name = "GPU 13 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, { .id = 0x10de0014, .name = "GPU 14 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, +{ .id = 0x10de0016, .name = "GPU 16 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, { .id = 0x10de0018, .name = "GPU 18 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, { .id = 0x10de0019, .name = "GPU 19 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, { .id = 0x10de001a, .name = "GPU 1a HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, @@ -1676,6 +1677,7 @@ MODULE_ALIAS("snd-hda-codec-id:10de0012"); MODULE_ALIAS("snd-hda-codec-id:10de0013"); MODULE_ALIAS("snd-hda-codec-id:10de0014"); +MODULE_ALIAS("snd-hda-codec-id:10de0016"); MODULE_ALIAS("snd-hda-codec-id:10de0018"); MODULE_ALIAS("snd-hda-codec-id:10de0019"); MODULE_ALIAS("snd-hda-codec-id:10de001a");
I think other 5xx cards need new value also.
Regards,
--- Richard Samson samson.richard@gmail.com
Richard Samson wrote at Sunday, February 27, 2011 6:11 PM:
... I added a new entry in patch_hdmi.c to get sound working on my own 560 Ti card :
--- patch_hdmi.c 2011-02-15 22:29:51.000000000 +0100 +++ patch_hdmi_GTX-560Ti.c 2011-02-27 23:52:56.984613414 +0100
...
patch_nvhdmi_8ch_89 }, { .id = 0x10de0014, .name = "GPU 14 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, +{ .id = 0x10de0016, .name = "GPU 16 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, { .id = 0x10de0018, .name = "GPU 18 HDMI/DP", .patch =
...
Acked-By: Stephen Warren swarren@nvidia.com
Hmm. When I submitted that last patch, I think I looked at the wrong branch of our Windows drivers, and hence missed a couple IDs, and added a couple of IDs that won't be used. Once this is merged, I'll follow up with another patch to fix the other issues (or let me know if you just want me to provide a single patch with everything in it, Takashi).
Thanks Richard.
At Mon, 28 Feb 2011 09:13:29 -0800, Stephen Warren wrote:
Richard Samson wrote at Sunday, February 27, 2011 6:11 PM:
... I added a new entry in patch_hdmi.c to get sound working on my own 560 Ti card :
--- patch_hdmi.c 2011-02-15 22:29:51.000000000 +0100 +++ patch_hdmi_GTX-560Ti.c 2011-02-27 23:52:56.984613414 +0100
...
patch_nvhdmi_8ch_89 }, { .id = 0x10de0014, .name = "GPU 14 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, +{ .id = 0x10de0016, .name = "GPU 16 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, { .id = 0x10de0018, .name = "GPU 18 HDMI/DP", .patch =
...
Acked-By: Stephen Warren swarren@nvidia.com
Hmm. When I submitted that last patch, I think I looked at the wrong branch of our Windows drivers, and hence missed a couple IDs, and added a couple of IDs that won't be used. Once this is merged, I'll follow up with another patch to fix the other issues (or let me know if you just want me to provide a single patch with everything in it, Takashi).
I don't mind in either way as long as it's a simple patch. As I already asked Richard for re-submitting with his sign-off, we can wait for a while.
BTW, this kind of id addition patch can be still merged in 2.6.38 if the chips listed there are real ones.
thanks,
Takashi
Hello,
Thank you for your replies.
As Stephen Warren said, a patch with full support of 5xx cards will be more useful. I'll be able to write a new patch but I don't have the knowledge to get other IDs.
Richard Samson
2011/2/28 Takashi Iwai tiwai@suse.de:
At Mon, 28 Feb 2011 09:13:29 -0800, Stephen Warren wrote:
Richard Samson wrote at Sunday, February 27, 2011 6:11 PM:
... I added a new entry in patch_hdmi.c to get sound working on my own 560 Ti card :
--- patch_hdmi.c 2011-02-15 22:29:51.000000000 +0100 +++ patch_hdmi_GTX-560Ti.c 2011-02-27 23:52:56.984613414 +0100
...
patch_nvhdmi_8ch_89 }, { .id = 0x10de0014, .name = "GPU 14 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, +{ .id = 0x10de0016, .name = "GPU 16 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, { .id = 0x10de0018, .name = "GPU 18 HDMI/DP", .patch =
...
Acked-By: Stephen Warren swarren@nvidia.com
Hmm. When I submitted that last patch, I think I looked at the wrong branch of our Windows drivers, and hence missed a couple IDs, and added a couple of IDs that won't be used. Once this is merged, I'll follow up with another patch to fix the other issues (or let me know if you just want me to provide a single patch with everything in it, Takashi).
I don't mind in either way as long as it's a simple patch. As I already asked Richard for re-submitting with his sign-off, we can wait for a while.
BTW, this kind of id addition patch can be still merged in 2.6.38 if the chips listed there are real ones.
thanks,
Takashi
At Mon, 28 Feb 2011 18:56:20 +0100, samson.richard@gmail.com wrote:
Hello,
Thank you for your replies.
As Stephen Warren said, a patch with full support of 5xx cards will be more useful. I'll be able to write a new patch but I don't have the knowledge to get other IDs.
OK, Stephen, could you prepare the patch including all missing IDs?
thanks,
Takashi
Richard Samson
2011/2/28 Takashi Iwai tiwai@suse.de:
At Mon, 28 Feb 2011 09:13:29 -0800, Stephen Warren wrote:
Richard Samson wrote at Sunday, February 27, 2011 6:11 PM:
... I added a new entry in patch_hdmi.c to get sound working on my own 560 Ti card :
--- patch_hdmi.c 2011-02-15 22:29:51.000000000 +0100 +++ patch_hdmi_GTX-560Ti.c 2011-02-27 23:52:56.984613414 +0100
...
patch_nvhdmi_8ch_89 }, { .id = 0x10de0014, .name = "GPU 14 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, +{ .id = 0x10de0016, .name = "GPU 16 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, { .id = 0x10de0018, .name = "GPU 18 HDMI/DP", .patch =
...
Acked-By: Stephen Warren swarren@nvidia.com
Hmm. When I submitted that last patch, I think I looked at the wrong branch of our Windows drivers, and hence missed a couple IDs, and added a couple of IDs that won't be used. Once this is merged, I'll follow up with another patch to fix the other issues (or let me know if you just want me to provide a single patch with everything in it, Takashi).
I don't mind in either way as long as it's a simple patch. As I already asked Richard for re-submitting with his sign-off, we can wait for a while.
BTW, this kind of id addition patch can be still merged in 2.6.38 if the chips listed there are real ones.
thanks,
Takashi
Le lundi 28 février 2011 à 19:09 +0100, Takashi Iwai a écrit :
At Mon, 28 Feb 2011 18:56:20 +0100, samson.richard@gmail.com wrote:
Hello,
Thank you for your replies.
As Stephen Warren said, a patch with full support of 5xx cards will be more useful. I'll be able to write a new patch but I don't have the knowledge to get other IDs.
OK, Stephen, could you prepare the patch including all missing IDs?
thanks,
Takashi
Hello,
I wrote an new patch with three new codec IDs for Nvidia Fermi cards. I attended to use existing IDs and online technical specifications to get two others cards : - Nvidia GeForce GTX 470 : id = 0x10de0015 - Nvidia GeForce GTX 560 Ti : id = 0x10de0016 (works fine) - Nvidia GeForce GTX 570 : id = 0x10de0017
Stephen, could you confirm these new IDs are correct ?
This patch applies on 2.6.38-rc6 and trunk.
Signed-off-by: Richard Samson samson.richard@gmail.com
diff -up sound/pci/hda/patch_hdmi.c{.orig,} > ~/snd-codec-hda_patch-hdmi_Nvidia_5xx.patch --- sound/pci/hda/patch_hdmi.c.orig 2011-03-01 14:15:07.380860984 +0100 +++ sound/pci/hda/patch_hdmi.c 2011-03-01 15:44:16.836444133 +0100 @@ -1634,6 +1634,9 @@ static struct hda_codec_preset snd_hda_p { .id = 0x10de0012, .name = "GPU 12 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, { .id = 0x10de0013, .name = "GPU 13 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, { .id = 0x10de0014, .name = "GPU 14 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, +{ .id = 0x10de0015, .name = "GPU 15 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, +{ .id = 0x10de0016, .name = "GPU 16 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, +{ .id = 0x10de0017, .name = "GPU 17 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, { .id = 0x10de0018, .name = "GPU 18 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, { .id = 0x10de0019, .name = "GPU 19 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, { .id = 0x10de001a, .name = "GPU 1a HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, @@ -1676,6 +1679,9 @@ MODULE_ALIAS("snd-hda-codec-id:10de0011" MODULE_ALIAS("snd-hda-codec-id:10de0012"); MODULE_ALIAS("snd-hda-codec-id:10de0013"); MODULE_ALIAS("snd-hda-codec-id:10de0014"); +MODULE_ALIAS("snd-hda-codec-id:10de0015"); +MODULE_ALIAS("snd-hda-codec-id:10de0016"); +MODULE_ALIAS("snd-hda-codec-id:10de0017"); MODULE_ALIAS("snd-hda-codec-id:10de0018"); MODULE_ALIAS("snd-hda-codec-id:10de0019"); MODULE_ALIAS("snd-hda-codec-id:10de001a");
Thanks. Richard Samson
Richard Samson
2011/2/28 Takashi Iwai tiwai@suse.de:
At Mon, 28 Feb 2011 09:13:29 -0800, Stephen Warren wrote:
Richard Samson wrote at Sunday, February 27, 2011 6:11 PM:
... I added a new entry in patch_hdmi.c to get sound working on my own 560 Ti card :
--- patch_hdmi.c 2011-02-15 22:29:51.000000000 +0100 +++ patch_hdmi_GTX-560Ti.c 2011-02-27 23:52:56.984613414 +0100
...
patch_nvhdmi_8ch_89 }, { .id = 0x10de0014, .name = "GPU 14 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, +{ .id = 0x10de0016, .name = "GPU 16 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, { .id = 0x10de0018, .name = "GPU 18 HDMI/DP", .patch =
...
Acked-By: Stephen Warren swarren@nvidia.com
Hmm. When I submitted that last patch, I think I looked at the wrong branch of our Windows drivers, and hence missed a couple IDs, and added a couple of IDs that won't be used. Once this is merged, I'll follow up with another patch to fix the other issues (or let me know if you just want me to provide a single patch with everything in it, Takashi).
I don't mind in either way as long as it's a simple patch. As I already asked Richard for re-submitting with his sign-off, we can wait for a while.
BTW, this kind of id addition patch can be still merged in 2.6.38 if the chips listed there are real ones.
thanks,
Takashi
Richard Samson wrote at Tuesday, March 01, 2011 8:15 AM:
Le lundi 28 février 2011 à 19:09 +0100, Takashi Iwai a écrit :
At Mon, 28 Feb 2011 18:56:20 +0100, samson.richard@gmail.com wrote:
As Stephen Warren said, a patch with full support of 5xx cards will be more useful.
OK, Stephen, could you prepare the patch including all missing IDs?
I wrote an new patch with three new codec IDs for Nvidia Fermi cards. I attended to use existing IDs and online technical specifications to get two others cards :
- Nvidia GeForce GTX 470 : id = 0x10de0015
- Nvidia GeForce GTX 560 Ti : id = 0x10de0016 (works fine)
- Nvidia GeForce GTX 570 : id = 0x10de0017
They don't appear to match our Windows driver. Where did you get the values from? (0x17 isn't in the Windows driver at all. 0x15 shouldn't be a GTX 470)
Thanks.
Le mardi 01 mars 2011 à 08:40 -0800, Stephen Warren a écrit :
Richard Samson wrote at Tuesday, March 01, 2011 8:15 AM:
Le lundi 28 février 2011 à 19:09 +0100, Takashi Iwai a écrit :
At Mon, 28 Feb 2011 18:56:20 +0100, samson.richard@gmail.com wrote:
As Stephen Warren said, a patch with full support of 5xx cards will be more useful.
OK, Stephen, could you prepare the patch including all missing IDs?
I wrote an new patch with three new codec IDs for Nvidia Fermi cards. I attended to use existing IDs and online technical specifications to get two others cards :
- Nvidia GeForce GTX 470 : id = 0x10de0015
- Nvidia GeForce GTX 560 Ti : id = 0x10de0016 (works fine)
- Nvidia GeForce GTX 570 : id = 0x10de0017
They don't appear to match our Windows driver. Where did you get the values from? (0x17 isn't in the Windows driver at all. 0x15 shouldn't be a GTX 470)
Thanks.
I ask to a reseller, it seems a bad idea. If I can help, where are these values in the Windows driver ?
Thanks. Richard Samson
Richard Samson wrote at Tuesday, March 01, 2011 10:16 AM:
Le mardi 01 mars 2011 à 08:40 -0800, Stephen Warren a écrit :
Richard Samson wrote at Tuesday, March 01, 2011 8:15 AM:
Le lundi 28 février 2011 à 19:09 +0100, Takashi Iwai a écrit :
At Mon, 28 Feb 2011 18:56:20 +0100, samson.richard@gmail.com wrote:
As Stephen Warren said, a patch with full support of 5xx cards will be more useful.
OK, Stephen, could you prepare the patch including all missing IDs?
I wrote an new patch with three new codec IDs for Nvidia Fermi cards. I attended to use existing IDs and online technical specifications to get two others cards :
- Nvidia GeForce GTX 470 : id = 0x10de0015
- Nvidia GeForce GTX 560 Ti : id = 0x10de0016 (works fine)
- Nvidia GeForce GTX 570 : id = 0x10de0017
They don't appear to match our Windows driver. Where did you get the values from? (0x17 isn't in the Windows driver at all. 0x15 shouldn't be a GTX 470)
I ask to a reseller, it seems a bad idea.
The only way to find the values that I know of would be to plug the card in, run Linux, and look at the ALSA /proc files. (Perhaps there's some equivalent tool under Windows)
If I can help, where are these values in the Windows driver ?
They encoded into the binary. I doubt they're user-accessible.
So, I think the best way to proceed is for you to modify your patch to add IDs 0x15 and 0x16 only. I'm pretty sure ID 0x17 isn't assigned to anything. Let's not remove any NVIDIA IDs even though I'm pretty sure a couple aren't actually assigned; if I'm wrong, we'd just have to add them back later, and having them in won't harm anything.
Thanks.
At Tue, 1 Mar 2011 09:45:54 -0800, Stephen Warren wrote:
Richard Samson wrote at Tuesday, March 01, 2011 10:16 AM:
Le mardi 01 mars 2011 à 08:40 -0800, Stephen Warren a écrit :
Richard Samson wrote at Tuesday, March 01, 2011 8:15 AM:
Le lundi 28 février 2011 à 19:09 +0100, Takashi Iwai a écrit :
At Mon, 28 Feb 2011 18:56:20 +0100, samson.richard@gmail.com wrote:
As Stephen Warren said, a patch with full support of 5xx cards will be more useful.
OK, Stephen, could you prepare the patch including all missing IDs?
I wrote an new patch with three new codec IDs for Nvidia Fermi cards. I attended to use existing IDs and online technical specifications to get two others cards :
- Nvidia GeForce GTX 470 : id = 0x10de0015
- Nvidia GeForce GTX 560 Ti : id = 0x10de0016 (works fine)
- Nvidia GeForce GTX 570 : id = 0x10de0017
They don't appear to match our Windows driver. Where did you get the values from? (0x17 isn't in the Windows driver at all. 0x15 shouldn't be a GTX 470)
I ask to a reseller, it seems a bad idea.
The only way to find the values that I know of would be to plug the card in, run Linux, and look at the ALSA /proc files. (Perhaps there's some equivalent tool under Windows)
If I can help, where are these values in the Windows driver ?
They encoded into the binary. I doubt they're user-accessible.
So, I think the best way to proceed is for you to modify your patch to add IDs 0x15 and 0x16 only. I'm pretty sure ID 0x17 isn't assigned to anything. Let's not remove any NVIDIA IDs even though I'm pretty sure a couple aren't actually assigned; if I'm wrong, we'd just have to add them back later, and having them in won't harm anything.
OK, I modified Richard's patch to get rid of 0x17 and merged now to sound git tree.
thanks,
Takashi
Le jeudi 03 mars 2011 à 12:52 +0100, Takashi Iwai a écrit :
At Tue, 1 Mar 2011 09:45:54 -0800, Stephen Warren wrote:
Richard Samson wrote at Tuesday, March 01, 2011 10:16 AM:
Le mardi 01 mars 2011 à 08:40 -0800, Stephen Warren a écrit :
Richard Samson wrote at Tuesday, March 01, 2011 8:15 AM:
Le lundi 28 février 2011 à 19:09 +0100, Takashi Iwai a écrit :
At Mon, 28 Feb 2011 18:56:20 +0100, samson.richard@gmail.com wrote: > As Stephen Warren said, a patch with full support of 5xx cards will be > more useful.
OK, Stephen, could you prepare the patch including all missing IDs?
I wrote an new patch with three new codec IDs for Nvidia Fermi cards. I attended to use existing IDs and online technical specifications to get two others cards :
- Nvidia GeForce GTX 470 : id = 0x10de0015
- Nvidia GeForce GTX 560 Ti : id = 0x10de0016 (works fine)
- Nvidia GeForce GTX 570 : id = 0x10de0017
They don't appear to match our Windows driver. Where did you get the values from? (0x17 isn't in the Windows driver at all. 0x15 shouldn't be a GTX 470)
I ask to a reseller, it seems a bad idea.
The only way to find the values that I know of would be to plug the card in, run Linux, and look at the ALSA /proc files. (Perhaps there's some equivalent tool under Windows)
If I can help, where are these values in the Windows driver ?
They encoded into the binary. I doubt they're user-accessible.
So, I think the best way to proceed is for you to modify your patch to add IDs 0x15 and 0x16 only. I'm pretty sure ID 0x17 isn't assigned to anything. Let's not remove any NVIDIA IDs even though I'm pretty sure a couple aren't actually assigned; if I'm wrong, we'd just have to add them back later, and having them in won't harm anything.
OK, I modified Richard's patch to get rid of 0x17 and merged now to sound git tree.
thanks,
Takashi
Hello,
Thank you for this commit.
Richard Samson
At Mon, 28 Feb 2011 02:10:54 +0100, Richard Samson wrote:
Hello,
It's my first patch/post here, i'm not sure if this list is the right place to suggest this patch.
To follow up a previous commit on snd-hda : http://git.kernel.org/?p=linux/kernel/git/tiwai/sound-2.6.git;a=commitdiff;h...
I added a new entry in patch_hdmi.c to get sound working on my own 560 Ti card :
--- patch_hdmi.c 2011-02-15 22:29:51.000000000 +0100 +++ patch_hdmi_GTX-560Ti.c 2011-02-27 23:52:56.984613414 +0100 @@ -1634,6 +1634,7 @@ { .id = 0x10de0012, .name = "GPU 12 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, { .id = 0x10de0013, .name = "GPU 13 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, { .id = 0x10de0014, .name = "GPU 14 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, +{ .id = 0x10de0016, .name = "GPU 16 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, { .id = 0x10de0018, .name = "GPU 18 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, { .id = 0x10de0019, .name = "GPU 19 HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, { .id = 0x10de001a, .name = "GPU 1a HDMI/DP", .patch = patch_nvhdmi_8ch_89 }, @@ -1676,6 +1677,7 @@ MODULE_ALIAS("snd-hda-codec-id:10de0012"); MODULE_ALIAS("snd-hda-codec-id:10de0013"); MODULE_ALIAS("snd-hda-codec-id:10de0014"); +MODULE_ALIAS("snd-hda-codec-id:10de0016"); MODULE_ALIAS("snd-hda-codec-id:10de0018"); MODULE_ALIAS("snd-hda-codec-id:10de0019"); MODULE_ALIAS("snd-hda-codec-id:10de001a");
Richard, could you your sign-off for merging to the upstream? See $LINUX/Documentation/SubmittingPatches, "Sign your work" section.
thanks,
Takashi
participants (4)
-
Richard Samson -
samson.richard＠gmail.com
-
Stephen Warren -
Takashi Iwai