HDSP driver update

HDSP driver update

[Thomas Charbonnel]

Hi,

Here's the long awaited hdsp driver update, sorry for the delay. No file
included in this post (they're too big). Jaroslav or Takashi, you can
get them from here:
http://www.undata.org/~thomas/

hdsp.c is a replacement for the actual hdsp.c file in
alsa-kernel/pci/rme9652/

hdsp.h should be placed in alsa-kernel/include

hdsploader is the firmware loading tool. It can be included to
alsa-tools. The firmware files can be removed from
alsa-kernel/pci/rme9652/, they are included with the archive (the 4 of
them).

hdspconf is a fltk gui for the card, it can also be included to
alsa-tools.

The driver now depends on the hwdep interface, the snd-hdsp entry of the
alsa-kernel/core/Makefile should be edited to add this dependency.

This driver should solve issues with the rev11 firmwares and the
hdsp9652 cards.

The card won't work until the firmware is uploaded with the hdspload
tool. In case the firmware gets lost (e.g. power supply failure), just
try to access the card again and the firmware will get uploaded back
from kernel cached version, or run the loader again. Note that at this
point the card will be reset to its default values (your mixer settings
will be lost). 

Thomas




-------------------------------------------------------
This SF.net email is sponsored by: ValueWeb: 
Dedicated Hosting for just $79/mo with 500 GB of bandwidth! 
No other company gives more support or power for your dedicated server
http://click.atdmt.com/AFF/go/sdnxxaff00300020aff/direct/01/

Re: HDSP driver update

[Paul Davis]

>Here's the long awaited hdsp driver update, sorry for the delay. No file

and just to make it clear, yes, i'm in favor of this update. it
(sh|w)ouldn't matter anyway, but thomas has done some really great
work here. even so, one question:

>The card won't work until the firmware is uploaded with the hdspload
>tool. 

am i right in thinking that this means that the driver will load, but
the card won't actually work?

--p


-------------------------------------------------------
This SF.net email is sponsored by: ValueWeb: 
Dedicated Hosting for just $79/mo with 500 GB of bandwidth! 
No other company gives more support or power for your dedicated server
http://click.atdmt.com/AFF/go/sdnxxaff00300020aff/direct/01/

Re: HDSP driver update

[Thomas Charbonnel]

> >The card won't work until the firmware is uploaded with the hdspload
> >tool. 
> 
> am i right in thinking that this means that the driver will load, but
> the card won't actually work?

The driver will load but only the hwdep interface will be created.
The only way to interact with the card is through the defined hwdep
ioctls. After the firmware is uploaded, all the remaining interfaces are
created (midi, pcm, ctl), and the card is fully functional.




-------------------------------------------------------
This SF.net email is sponsored by: ValueWeb: 
Dedicated Hosting for just $79/mo with 500 GB of bandwidth! 
No other company gives more support or power for your dedicated server
http://click.atdmt.com/AFF/go/sdnxxaff00300020aff/direct/01/

Re: HDSP driver update

[Takashi Iwai]

Hi,

At 07 Apr 2003 15:07:13 +0200,
Thomas Charbonnel wrote:
> 
> 
> > >The card won't work until the firmware is uploaded with the hdspload
> > >tool. 
> > 
> > am i right in thinking that this means that the driver will load, but
> > the card won't actually work?
> 
> The driver will load but only the hwdep interface will be created.
> The only way to interact with the card is through the defined hwdep
> ioctls. After the firmware is uploaded, all the remaining interfaces are
> created (midi, pcm, ctl), and the card is fully functional.

that's nice.  so you succeeded dieting :)

i committed the new driver and hdsploader to cvs tree right now.
i had to fix the compilation error in hdsp.c because of the last
change of control interface (inside the kernel).

the two issues still pending:

1. hdspconf is still not merged yet to cvs, because it's written in
   C++ and fltk.  i'm not sure whether we should put it into the same
   tree.  IMO, it's ok - we already have envy24control with gtk.

   so, if no one has objection, i'll try to put to alsa-tools, too.


2. the change of control interface above would make the access to a
   matrix mixer over the standard control API much easier without
   using the non-official indirect accessing.

   could you try to port the driver to use that feature?


thanks,

Takashi


-------------------------------------------------------
This SF.net email is sponsored by: ValueWeb: 
Dedicated Hosting for just $79/mo with 500 GB of bandwidth! 
No other company gives more support or power for your dedicated server
http://click.atdmt.com/AFF/go/sdnxxaff00300020aff/direct/01/

Re: HDSP driver update

[Thomas Charbonnel]

[-- Attachment #1: Type: text/plain, Size: 986 bytes --]


> i committed the new driver and hdsploader to cvs tree right now.
> i had to fix the compilation error in hdsp.c because of the last
> change of control interface (inside the kernel).
> 
Thanks

> the two issues still pending:
> 
> 1. hdspconf is still not merged yet to cvs, because it's written in
>    C++ and fltk.  i'm not sure whether we should put it into the same
>    tree.  IMO, it's ok - we already have envy24control with gtk.
> 
>    so, if no one has objection, i'll try to put to alsa-tools, too.
> 
Just FYI the upcoming hdspmixer you wanted to include too is also
C++/fltk.

> 
> 2. the change of control interface above would make the access to a
>    matrix mixer over the standard control API much easier without
>    using the non-official indirect accessing.
> 
>    could you try to port the driver to use that feature?
> 
> 
Of course.

The attached patch fixes a problem with hwdep interface creation in the
case a firmware is detected at boot time.

Thomas


[-- Attachment #2: hdsp_hwdep_create.patch --]
[-- Type: text/plain, Size: 325 bytes --]

--- hdsp.c.old	2003-04-07 17:32:36.000000000 +0200
+++ hdsp.c	2003-04-08 11:05:22.000000000 +0200
@@ -4116,6 +4116,10 @@
 			hdsp->io_type = Digiface;
 		}
 		
+		if ((err = snd_hdsp_create_hwdep(card, hdsp)) < 0) {
+			return err;
+		}
+		
 		snd_hdsp_initialize_channels(hdsp);
 		
 		snd_hdsp_initialize_midi_flush(hdsp);

Re: HDSP driver update

[Takashi Iwai]

Hi Thomas,

At 08 Apr 2003 11:15:25 +0200,
Thomas Charbonnel wrote:
> 
> > the two issues still pending:
> > 
> > 1. hdspconf is still not merged yet to cvs, because it's written in
> >    C++ and fltk.  i'm not sure whether we should put it into the same
> >    tree.  IMO, it's ok - we already have envy24control with gtk.
> > 
> >    so, if no one has objection, i'll try to put to alsa-tools, too.
> > 
> Just FYI the upcoming hdspmixer you wanted to include too is also
> C++/fltk.

ok.

> 
> > 
> > 2. the change of control interface above would make the access to a
> >    matrix mixer over the standard control API much easier without
> >    using the non-official indirect accessing.
> > 
> >    could you try to port the driver to use that feature?
> > 
> > 
> Of course.
> 
> The attached patch fixes a problem with hwdep interface creation in the
> case a firmware is detected at boot time.

thanks, applied.

i also fixed the old style of initialization of structs in hdsp.c.
please update your cvs tree.


ciao,

Takashi


-------------------------------------------------------
This SF.net email is sponsored by: ValueWeb: 
Dedicated Hosting for just $79/mo with 500 GB of bandwidth! 
No other company gives more support or power for your dedicated server
http://click.atdmt.com/AFF/go/sdnxxaff00300020aff/direct/01/