Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 18 Apr 2002 22:26:57 +0100
From:      jmdupoux@lineone.net
To:        John Utz <john@utzweb.net>
Cc:        questions@freebsd.org, <freebsd-multimedia@freebsd.org>
Subject:   Re: Kernel config parameters for soundcard - 4.4 release
Message-ID:  <3CBF4831.25173.EF10C7@localhost>
In-Reply-To: <Pine.LNX.4.44.0204181430010.12819-100000@jupiter.linuxengine.net>
References:  <3CBF10E9.5DF9F28D@lineone.net>

next in thread | previous in thread | raw e-mail | index | archive | help
hmmm, I really should  have mentioned this in my last mail as it is the obvious thing 
to suggest (how often  do we overlook things like that when we get too wrapped up 
in a problem), but  I already tried  compiling with  just:

device pcm
device  sb0 at isa? port 0x220 irq 5 drq 0 flags 0x11

and none of the other devices, but  make  still falls over with the exact same errors.
I think I tried  more or less every possible combination of inclusion/exclusion of the 
other devices combined with  pcm and  sb0, but  nothing  ever got past  make   ( so 
long as I had  pcm and not snd as the first device)

Since you're  working on the code for this I'll be more specific  about the exact card 
and chipsets involved here as it might be useful for reference:

Aztech Sound Galaxy pro II (often  referred to  as Sound Galaxy Washington or 
sometimes Nova)
main chipset is AZT2316r  and the card  also  has a Crystal  4231a-kl

obviously the card has  DOS/Windows drivers, which basically set  the card up with 
SB pro emulation, combined with MSS sound system,  opl3,  joystick and midi, all 
at the "standard"  addresses (software adjustable)
I think I also tried to install Crystal drivers on the card once in DOS but definitely 
without very good results.
Other non-specific DOS programs  can get audio produced  using  basic  SB-pro 
settings without the specific Aztech  driver (eg a  DOS general hardware 
diagnostics utility I use to check IRQ's )
From what I can tell the card is not truly  plug-n-play, but uses  an EEPROM chip 
where the settings  get written to when options get changed through the driver 
interface


I'm not  desparate  to get this working although it would be very useful, but still very  
confused why the __SGNXPRO__ option doesn't  get  recognised -  this  is where 
practical  experience from someone else's  machine might really pay off.


Jean-Mark Dupoux
jmdupoux_@_lineone.net



On 18 Apr 2002 at 14:55, John Utz wrote:

> midi is utterly unimplemented with the new sound code.
> 
> so putting the midi devices into the config file will generate the errors 
> you are getting, because the old midi code is based on 'oldbus' for want 
> of a better term, whereas the device structure in 4.X is based on 
> 'newbus'.
> 
> so that stuff aint gonna compile.
> 
> i am *actively* working on merging the tanimura code into 4.5-RELEASE. i 
> spend 2+ hours a day on it, 5 days a week ( assuming that i take the bus 
> to work all week, if i drive to work, then i dont work on it that day, eh! 
> :-) ).
> 
> i've been thinking about making a public announcement that i was going to 
> tackle this, and i'd rather have waited until i had working code ( which 
> i am guessing is 2 months off ), but your mail seemed like a good segue.
> 
> the patches that tanimura generated in 1999 are pretty big, and so i have 
> been basically wading thru the code trying to understand the 
> relationships.
> 
> my intent is to turn the midi bits into a kld. that makes it a lot harder 
> in the beginning ( because the code *appears* to be implemented as 
> compiled in only...but i could be mistaken ), but makes development easier 
> in the long run because the build/try cycle is much shorter ( i usually 
> refer to this as 'compile, fiddle, compile'  :-) ).
> 
> streamlining the process is *very* important to me because i am doing this 
> on an HP 800CT laptop with a p166MMX and 80M of ram. we ain't talkin' big 
> iron here.....
> 
> so, the good news for you is that the 800CT has an ESS AudioDrive ISA OEM 
> in it, so it too makes use of the sbc bridge. so any work that i do should 
> be immediately available to you.
> 
> after i get it to work reliably on my laptop, i'll work on the other 
> hardware that i have at home, including my real live mpu-401 card.
> 
> it's important to note that i make no predictions as to when i will 
> finish or if i will even be successful....midi has a long sad story with 
> freebsd because it's *never* been finished! i've been running freebsd 
> since pre 1.0 and i have seen various bits of midi code thru the years. 
> individuals take up the cause, and then they stop, and are never heard 
> from again! lord knows that i have sent tanimura-san an email or two since 
> 1999 and i have never heard bit0 from him. or her.
> 
> i *do* get lots of mail that sez 'gee, it would make me really happy if i 
> could use midi with freebsd'.
> 
> but the code is pretty dense, and FreeBSD kernel-hackery seems to have an 
> awfully steep learning curve, it's certainly kicked *my* ass more than 
> once.
> 
> as soon as i get something to compile, i'll post to -multimedia. then we 
> can go from there.....
> 
> have a nice day!
> 
> johnu
> 
> 
> On Thu, 18 Apr 2002, Jean-Mark Dupoux wrote:
> 
> > The first thing I tried  was running the compile with  pcm instead of
> > snd, as per all the docs but then I got  different errors  come up, this
> > time during the make.  That's the only reason  I compiled with snd
> > instead. 
> > 
> > Specifically, I get a screenful of errors similar to the 2 or 3 lines
> > i've quoted below, and then the build stops:
> > 
> > midi_synth.o: In function 'midi_synth_xxxxx':
> > midi_synth.o(.text+0xNNN): undefined reference to 'synth_devs'
> > /* (similar error definitions complaining about other midi.synth.o
> > function calls and undefined  references repeated for several lines)  */
> > 
> > ending with the line:
> > 
> > *** Error code 1
> > 
> > and an abrupt stop to the make.
> > 
> > 
> > But in any case, it wasnt until I had a "successful" build that I
> > realised no mixer was being  created, so then I went back and tried to
> > compile with pcm again instead of snd and this time include the line
> > with options __SGNXPRO__ added for the mixer, but then the build  falls
> > over when I run /usr/sbin/config MYKERNELFILE with the exact same  error
> > "Unknown option __SGNXPRO__"
> > 
> > still cant make sense of this one, all the other documented options
> > i've  read about for the sb device dont kick up any errors (but they're 
> > not  relevant to my card !)
> > 
> > 
> > thanks in advance
> > 
> > John Utz wrote:
> > > 
> > > why are you using snd instead of pcm?
> > > 
> > > snd is pretty deprecated....
> > > 
> > > i think you will get immediately better results using sbc ( the sb bridge
> > > driver for pcm ) and pcm
> > > 
> > > man pcm and man sbc can be your new best friends!
> > > 
> > > On Thu, 18 Apr 2002, Jean-Mark Dupoux wrote:
> > > 
> > > > Im having  problems with an option for soundcard  support to put in a
> > > > kernel config file. To begin with I added the following lines into the
> > > > config file:
> > > >
> > > > device snd
> > > > device sb0 at isa? port 0x220 irq 5 drq 0 flags 0x11
> > > > device sbmidi0 at isa? port 0x330 irq 9
> > > > device opl0 at isa? port 0x388
> > > > device joy0 at isa? port 0x200
> > > >
> > > > The kernel compiled properly and successfully loaded the soundblaster
> > > > pro 3.1  driver. But I still got  no sound out of the card, using cat
> > > > /filename > /dev/audio or > /dev/dsp  and got the same no  sound
> > > > result  using play /filename
> > > >
> > > > when I checked  cat /dev/sndstat it showed that no mixer was loaded,
> > > > which seemed like a fair explanation for the silence, so I dug  deeper
> > > > in man sb(4). This turned up an extra option for the Sound Galaxy pro
> > > > card that I am using which should specifically add mixer support - the
> > > > SG card  apparently has a slightly different  mixer from a basic  SB pro
> > > > which  includes bass/treble adjustment, and I hoped the problem would be
> > > > solved.
> > > > So I went ahead and added the suggestion from man sb
> > > >
> > > > options __SGNXPRO__
> > > >
> > > > into the config file, right after the sound devices  section, but  now
> > > > the  kernel doesnt compile any more. After running
> > > > ln -s /usr/local/kernels/MYKERNELFILE
> > > > I then enter as usual:
> > > > /usr/sbin/config MYKERNELFILE
> > > >
> > > > but that's  as far as the process gets. The output on the screen is:
> > > > Error
> > > > Unknown option  __SGNXPRO__ (line xxx)
> > > >
> > > > I've  checked and double-checked the relevant  documentation and im sure
> > > > I havnt made a typo or syntax error.
> > > > What's more that specific  option seems  to have been present going back
> > > > quite  a few releases, so I'm truly puzzled why config  wont  accept it
> > > >
> > > > Has anyone else  got experience with this kind of card, or with
> > > > options  parameters for soundcards generally, or should I be thinking
> > > > about reporting this to bugs ?
> > > >
> > > > thanks in advance
> > > > Jean-Mark Dupoux
> > > > jmdupoux_@_lineone.net
> > > >
> > > > To Unsubscribe: send mail to majordomo@FreeBSD.org
> > > > with "unsubscribe freebsd-multimedia" in the body of the message
> > > >
> > > 
> > > --
> > > 
> > > John L. Utz III
> > > john@utzweb.net
> > > 
> > > Idiocy is the Impulse Function in the Convolution of Life
> > 
> 
> -- 
> 
> John L. Utz III
> john@utzweb.net
> 
> Idiocy is the Impulse Function in the Convolution of Life
> 



To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-questions" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3CBF4831.25173.EF10C7>