HDA sound driver mod for sigmatel 92xx

[Eric Anderson]

On 07/13/06 05:05, Boris Samorodov wrote:
> On Wed, 12 Jul 2006 22:34:05 -0500 Eric Anderson wrote:
> 
>> On 07/12/06 20:46, Wesley Morgan wrote:
>>> On Wed, 12 Jul 2006, Eric Anderson wrote:
>>>
>>>> On 07/12/06 20:04, Wesley Morgan wrote:
>>>>> On Wed, 12 Jul 2006, Eric Anderson wrote:
>>>>>
>>>>>> On 07/09/06 09:11, Wesley Morgan wrote:
>>>>>>> With a little help from the netbsd azalia driver (a wonderful
>>>>>>> piece of work, by the way) and the hda specs, I've been able to
>>>>>>> get Andrea's hdac code to work with my laptop. I believe the
>>>>>>> problem was that the codec and dac started out in some kind of
>>>>>>> low-power mode and needed to be woken up before they would work.
>>>>>>>
>>>>>>> Hopefully someone is still working on a proper driver, but for
>>>>>>> now, I can watch movies again!
>>>>>>>
>>>>>>> A diff against Andrea's work is attached.
>>>>>>
>>>>>> I couldn't get this patch to work - all hunks fail.  What am I
>>>>>> doing wrong?
>>>>> Make sure your mailer didn't wrap long lines? It applies cleanly
>>>>> for me. I've attached a newer diff that makes the mixer work
>>>>> properly as well, but it's hard-coded to the correct widget for me
>>>>> so it might not work for everyone.
>>>> Well, it must be my hdac.tgz then, because the attached patch looks
>>>> fine, but still fails poorly.  Maybe a pointer to the hdac.tgz you
>>>> used, and an md5 sum would do the trick..
>>> The original hdac.tgz posted by Andrea, I believe can be found at:
>>>
>>> http://darkircop.org/hdac.tgz
>>>
>>>
> 
>> [root at neutrino /tmp/freebsd]# fetch http://darkircop.org/hdac.tgz
>> hdac.tgz                                      100% of  122 kB  151 kBps
>> [root at neutrino /tmp/freebsd]# tar xvzf hdac.tgz
>> x hdac
>> x hdac/hdac.c
>> x hdac/hdac_reg.h
>> x hdac/hdac.h
>> x hdac/hda_reg.h
>> x hdac/hdac_private.h
>> x hdac/Makefile
>> x hdac/hdac.z~
>> x hdac/calc.core
>> x hdac/hdac.y~
>> x hdac/hdac.c~
>> x hdac/hdac.c.bak
>> x hdac/.hda_reg.h.swp
>> [root at neutrino /tmp/freebsd]# cd hdac
>> [root at neutrino /tmp/freebsd/hdac]# patch < ../hdac.diff
>> Hmm...  Looks like a unified diff to me...
>> The text leading up to this was:
>> --------------------------
>> |diff -urb hdac/hdac.c hdac-sigma/hdac.c
>> |--- hdac/hdac.c        Mon Jun  5 15:15:12 2006
>> |+++ hdac-sigma/hdac.c  Wed Jul 12 21:02:32 2006
>> --------------------------
>> Patching file hdac.c using Plan A...
>> Hunk #1 failed at 247.
>> Hunk #2 failed at 261.
>> Hunk #3 succeeded at 1879 with fuzz 2 (offset 1605 lines).
>> Hunk #4 failed at 2396.
>> Hunk #5 failed at 2430.
>> Hunk #6 failed at 2472.
>> Hunk #7 failed at 2517.
>> Hunk #8 failed at 2568.
>> Hunk #9 failed at 2597.
>> Hunk #10 failed at 2893.
>> Hunk #11 failed at 3233.
>> 10 out of 11 hunks failed--saving rejects to hdac.c.rej
>> done
>> [root at neutrino /tmp/freebsd/hdac]#
> 
> Just in case: make sure you don't have CRLFs at line endings of your
> patch.
> 
> 
> WBR


*sigh* - that was it.  How's they get introduced?  My mailer didn't do 
it (I just checked).  Must have been on the sending side?

Painful.

Only one hunk failed that way:

[root at neutrino /tmp/freebsd]# tar xvzf hdac.tgz
x hdac
x hdac/hdac.c
x hdac/hdac_reg.h
x hdac/hdac.h
x hdac/hda_reg.h
x hdac/hdac_private.h
x hdac/Makefile
x hdac/hdac.z~
x hdac/calc.core
x hdac/hdac.y~
x hdac/hdac.c~
x hdac/hdac.c.bak
x hdac/.hda_reg.h.swp
[root at neutrino /tmp/freebsd]# cd hdac
[root at neutrino /tmp/freebsd/hdac]# patch <../hdac-sigma2.diff
Hmm...  Looks like a unified diff to me...
The text leading up to this was:
--------------------------
|diff -urb hdac/hdac.c hdac-sigma/hdac.c
|--- hdac/hdac.c        Mon Jun  5 15:15:12 2006
|+++ hdac-sigma/hdac.c  Wed Jul 12 21:02:32 2006
--------------------------
Patching file hdac.c using Plan A...
Hunk #1 succeeded at 247.
Hunk #2 succeeded at 261.
Hunk #3 succeeded at 274.
Hunk #4 failed at 791.
Hunk #5 succeeded at 825.
Hunk #6 succeeded at 867.
Hunk #7 succeeded at 912.
Hunk #8 succeeded at 963.
Hunk #9 succeeded at 992.
Hunk #10 succeeded at 1288.
Hunk #11 succeeded at 1628.
1 out of 11 hunks failed--saving rejects to hdac.c.rej
done
[root at neutrino /tmp/freebsd/hdac]# cat hdac.c.rej
***************
*** 786,791 ****
         devinfo->stepping_id = 
HDA_PARAM_REVISION_ID_STEPPING_ID(revisionid);
         devinfo->node_type = HDA_PARAM_FCT_GRP_TYPE_NODE_TYPE(fctgrptype);

         hdac_add_child(sc, devinfo);
   }

--- 791,800 ----
         devinfo->stepping_id = 
HDA_PARAM_REVISION_ID_STEPPING_ID(revisionid);
         devinfo->node_type = HDA_PARAM_FCT_GRP_TYPE_NODE_TYPE(fctgrptype);

+
+       device_printf(sc->dev, "Vendor info: %x %x %x %x %x %x\n", 
vendorid, devinfo->vendor_id, devinfo->device_id,
+               devinfo->revision_id, devinfo->stepping_id, 
devinfo->node_type);
+
         hdac_add_child(sc, devinfo);
   }



After manually applying the patch, it builds, and makes my sound work 
fine!!!!!!

Woohoo!!

Thanks to those who made this work!

Eric


-- 
------------------------------------------------------------------------
Eric Anderson        Sr. Systems Administrator        Centaur Technology
Anything that works is better than anything that doesn't.
------------------------------------------------------------------------