[PATCH] Tascam US-224/428 MMC Implementation

[PATCH] Tascam US-224/428 MMC Implementation

[Rui Nuno Capela]

[-- Attachment #1: Type: text/plain, Size: 1451 bytes --]

Hi everybody,

This is my latest patch against the alsa-tools tree (as of todays CVS
HEAD), regarding the us428control MMC implementation as for the Tascam
US-224/428 audio/midi USB control surfaces:

    us428control now bumps to version 0.4.4.

Interesting points are: jog wheel dialing is now alternating between
spitting out some MMC-Step or MMC-Shuttle messages, depending on transport
state, instead of those MMC-Locate(s) which was an early implementation
mistake of mine.

Controlling ardour with (my) US-224 is getting much more fun now :)

However, there's still some little annoyances due to my lack of advice
regarding ardour's Play/Record control behaviour:

1) Ardour doesn't seem to send out MMC-Play messages when one starts
transport within it (e.g. by clicking the playback button widget).

2) Ardour doesn't seem to react against MMC-RecordPause, which I assume
(probably erroneously) it would be equivalent to clicking on its record
widget button.

3) Sending a MMC-RecordStrobe message to ardour has the strange effect to
start playback while toggling recording off instantaneously.

AFAICT these issues seems to be specific to ardour (0.9beta18.4), not to
us428control which is just sending out the proper MMC messages. Or that I
think.

I would like to settle all this before I'm going off on holidays. Hope
it's welcome and on time for the imminent alsa dot-6 release :)

Cheers,
-- 
rncbc aka Rui Nuno Capela
rncbc@rncbc.org


[-- Attachment #2: us428control-0.4.4.patch.gz --]
[-- Type: application/x-gzip, Size: 2145 bytes --]

Re: [PATCH] Tascam US-224/428 MMC Implementation

[Takashi Iwai]

At Sun, 1 Aug 2004 21:15:04 +0100 (WEST),
Rui Nuno Capela wrote:
> 
> Hi everybody,
> 
> This is my latest patch against the alsa-tools tree (as of todays CVS
> HEAD), regarding the us428control MMC implementation as for the Tascam
> US-224/428 audio/midi USB control surfaces:
> 
>     us428control now bumps to version 0.4.4.

Patch applied.  Thanks!


Takashi


-------------------------------------------------------
This SF.Net email is sponsored by OSTG. Have you noticed the changes on
Linux.com, ITManagersJournal and NewsForge in the past few weeks? Now,
one more big change to announce. We are now OSTG- Open Source Technology
Group. Come see the changes on the new OSTG site. www.ostg.com

Re: [PATCH] Tascam US-224/428 MMC Implementation

[Rui Nuno Capela]

Hi Takashi,

> At Sun, 1 Aug 2004 21:15:04 +0100 (WEST),
> Rui Nuno Capela wrote:
>>
>> Hi everybody,
>>
>> This is my latest patch against the alsa-tools tree (as of todays CVS
>> HEAD), regarding the us428control MMC implementation as for the Tascam
>> US-224/428 audio/midi USB control surfaces:
>>
>>     us428control now bumps to version 0.4.4.
>
> Patch applied.  Thanks!
>
> Takashi

BTW, I think this has been asked before but didn't had the chance to know
the right answer, so what's keeping snd-usb-usx2y out of the stock 2.6
kernel.org tree?

Anything to do with firmware licensing? IIRC there's already a GPL'd
bootstrap loader...

Regards.
-- 
rncbc aka Rui Nuno Capela
rncbc@rncbc.org



-------------------------------------------------------
This SF.Net email is sponsored by OSTG. Have you noticed the changes on
Linux.com, ITManagersJournal and NewsForge in the past few weeks? Now,
one more big change to announce. We are now OSTG- Open Source Technology
Group. Come see the changes on the new OSTG site. www.ostg.com

Re: [PATCH] Tascam US-224/428 MMC Implementation

[Takashi Iwai]

At Mon, 2 Aug 2004 13:12:15 +0100 (WEST),
Rui Nuno Capela wrote:
> 
> Hi Takashi,
> 
> > At Sun, 1 Aug 2004 21:15:04 +0100 (WEST),
> > Rui Nuno Capela wrote:
> >>
> >> Hi everybody,
> >>
> >> This is my latest patch against the alsa-tools tree (as of todays CVS
> >> HEAD), regarding the us428control MMC implementation as for the Tascam
> >> US-224/428 audio/midi USB control surfaces:
> >>
> >>     us428control now bumps to version 0.4.4.
> >
> > Patch applied.  Thanks!
> >
> > Takashi
> 
> BTW, I think this has been asked before but didn't had the chance to know
> the right answer, so what's keeping snd-usb-usx2y out of the stock 2.6
> kernel.org tree?
> 
> Anything to do with firmware licensing? IIRC there's already a GPL'd
> bootstrap loader...

No, just because I don't know whether the driver works on 2.6 kernel.
No one reported/requested on alsa-devel ML yet.

If it's confirmed to work, it's ok to push it up to 2.6 tree after
some reviews/clean-up's.


Takashi


-------------------------------------------------------
This SF.Net email is sponsored by OSTG. Have you noticed the changes on
Linux.com, ITManagersJournal and NewsForge in the past few weeks? Now,
one more big change to announce. We are now OSTG- Open Source Technology
Group. Come see the changes on the new OSTG site. www.ostg.com

Re: [PATCH] Tascam US-224/428 MMC Implementation

[Rui Nuno Capela]

Hi Takashi, Karten, Martin,
everyone else,

>> BTW, I think this has been asked before but didn't had the chance to
>> know the right answer, so what's keeping snd-usb-usx2y out of the
>> stock 2.6 kernel.org tree?
>>
>> Anything to do with firmware licensing? IIRC there's already a GPL'd
>> bootstrap loader...
>
> No, just because I don't know whether the driver works on 2.6 kernel.
> No one reported/requested on alsa-devel ML yet.
>
> If it's confirmed to work, it's ok to push it up to 2.6 tree after
> some reviews/clean-up's.
>

OK. Now that you're asking, I've been testing on a couple of my boxes
and the results are a mix of success and a complete disappointment:

1) SUSE 9.1 Professional, vanilla 2.6.8-rc2 SMP/HT preemptible, NPTL 0.61:
   - hotplug/firmware load: OK.
     - snd-usb-usx2y 1.0.5a: OK, jackd RT: OK.
     - snd-usb-usx2y CVS HEAD 2004-08-01: OK, jackd RT: OK.

2) Mandrake 10.0 Official vanilla 2.6.8-rc2 UP preemptible, NPTL 0.60:
   - hotplug/firmware load: OK.
     - snd-usb-usx2y 1.0.5a: OK, jackd RT: unstable, non-RT seems OK.
     - snd-usb-usx2y CVS HEAD 2004-08-01: KO. HORROR!
       As soon as the pcm device is requested (e.g. via aplay)
       the whole system freezes, puff!
     This is a true showstopper as a matter of fact. FWIW it doesn't
     matter whether on NPTL or linuxthreads. In fact, I don't even know
     if this pthreads stuff has any influence at all. Certainly not.
     Please note that last patch of Karsten's probably did the
     difference, unfortunately for worse :/

FYI I'm on 2.6 since January, and I can certainly argue that on this
SUSE box everything is working flawlessly, period. Problem comes on that
Mdk box, don't know why. Is it due to SMP/HT? OTOH I don't even know if
all this thing works on 2.4 at all :) but I'm the only with a US-224 here,
am I?

Other US-x2y users may speak freely :)

Cheers.
-- 
rncbc aka Rui Nuno Capela
rncbc@rncbc.org



-------------------------------------------------------
This SF.Net email is sponsored by OSTG. Have you noticed the changes on
Linux.com, ITManagersJournal and NewsForge in the past few weeks? Now,
one more big change to announce. We are now OSTG- Open Source Technology
Group. Come see the changes on the new OSTG site. www.ostg.com

Re: [PATCH] Tascam US-224/428 MMC Implementation

[Takashi Iwai]

At Mon, 2 Aug 2004 16:43:09 +0100 (WEST),
Rui Nuno Capela wrote:
> 
> Hi Takashi, Karten, Martin,
> everyone else,
> 
> >> BTW, I think this has been asked before but didn't had the chance to
> >> know the right answer, so what's keeping snd-usb-usx2y out of the
> >> stock 2.6 kernel.org tree?
> >>
> >> Anything to do with firmware licensing? IIRC there's already a GPL'd
> >> bootstrap loader...
> >
> > No, just because I don't know whether the driver works on 2.6 kernel.
> > No one reported/requested on alsa-devel ML yet.
> >
> > If it's confirmed to work, it's ok to push it up to 2.6 tree after
> > some reviews/clean-up's.
> >
> 
> OK. Now that you're asking, I've been testing on a couple of my boxes
> and the results are a mix of success and a complete disappointment:
> 
> 1) SUSE 9.1 Professional, vanilla 2.6.8-rc2 SMP/HT preemptible, NPTL 0.61:
>    - hotplug/firmware load: OK.
>      - snd-usb-usx2y 1.0.5a: OK, jackd RT: OK.
>      - snd-usb-usx2y CVS HEAD 2004-08-01: OK, jackd RT: OK.
> 
> 2) Mandrake 10.0 Official vanilla 2.6.8-rc2 UP preemptible, NPTL 0.60:
>    - hotplug/firmware load: OK.
>      - snd-usb-usx2y 1.0.5a: OK, jackd RT: unstable, non-RT seems OK.
>      - snd-usb-usx2y CVS HEAD 2004-08-01: KO. HORROR!
>        As soon as the pcm device is requested (e.g. via aplay)
>        the whole system freezes, puff!

That's bad.  Can you get any trace via magic sysrq?

>      This is a true showstopper as a matter of fact. FWIW it doesn't
>      matter whether on NPTL or linuxthreads. In fact, I don't even know
>      if this pthreads stuff has any influence at all. Certainly not.
>      Please note that last patch of Karsten's probably did the
>      difference, unfortunately for worse :/
> 
> FYI I'm on 2.6 since January, and I can certainly argue that on this
> SUSE box everything is working flawlessly, period. Problem comes on that
> Mdk box, don't know why. Is it due to SMP/HT?

Could be.  Can you try a UP kernel on Mandrake for testing?


Anyway, we'll need clean up of codes before moving to alsa-kernel
tree.  The all 2.4/2.2-specific codes should be stripped down and
provided as patches in alsa-driver.

I can do it but I'm afraid of breaking something.  Could any usx2y
owner take this action?


Takashi


-------------------------------------------------------
This SF.Net email is sponsored by OSTG. Have you noticed the changes on
Linux.com, ITManagersJournal and NewsForge in the past few weeks? Now,
one more big change to announce. We are now OSTG- Open Source Technology
Group. Come see the changes on the new OSTG site. www.ostg.com

Re: [PATCH] Tascam US-224/428 MMC Implementation

[Takashi Iwai]

[-- Attachment #1: Type: text/plain, Size: 1073 bytes --]

At Mon, 02 Aug 2004 17:52:35 +0200,
I wrote:
> 
> Anyway, we'll need clean up of codes before moving to alsa-kernel
> tree.  The all 2.4/2.2-specific codes should be stripped down and
> provided as patches in alsa-driver.
> 
> I can do it but I'm afraid of breaking something.  Could any usx2y
> owner take this action?

Ok, since no one gave his name, I did it :)


Now it's moved to alsa-kernel CVS tree.

Along with this action, the following items have been changed:

- the codes for old kernels are patched.  usbusx2y.c and usX2Yhwdep.c
  have the patches.

- most of snd_printd() is replaced with snd_printdd().
  CONFIG_SND_DEBUG is used commonly, and it must be annoying to print
  kernel messages at each time you start PCM.

- the complete handlers are moved to usbusx2.c from usX2Yhwdep.c
  to simplify the patches.  Instead, the initializers are exported.

- __cplusplus codes are removed from the header.

- Karsten's last patch to 0.7.2 is already merged.


I attached the patches below, so that you can test immediately before
sf's tree is synced.


Takashi

[-- Attachment #2: usx2y-kernel.dif.gz --]
[-- Type: application/octet-stream, Size: 16104 bytes --]

[-- Attachment #3: usx2y-driver.dif.gz --]
[-- Type: application/octet-stream, Size: 19776 bytes --]