Re: pcm_chn_create failed with 2004-08-12 kernel

On 2004-08-13T09:02:31+0900, Taku YAMAMOTO wrote:
> On Thu, 12 Aug 2004 15:53:54 -0700 (PDT)
> Don Lewis <truckman_at_FreeBSD.org> wrote:
> > On 12 Aug, Michael W. Oliver wrote:
> > > Ok, this world was cvsup'd on 2004.08.12 at 09:59:03 EDT.  I removed the
> > > 'sound' and 'snd_maestro' device lines from my kernel config and decided
> > > to load them as modules (after I got the same output shown below with
> > > sound and snd_maestro compiled into the kernel).  When I kldload
> > > 'sound', I get no output.  After that, when I kldload 'snd_maestro', I
> > > get the following output:
> > > 
> > > 
> > > pcm0: <ESS Technology Maestro-2E> port 0x1400-0x14ff irq 10 at device 8.0 on pci0
> > > pcm0: <SigmaTel STAC9721/23 AC97 Codec>
> > > pcm0: [GIANT-LOCKED]
> > > pcm0: offset 0xfffa1000 exceeds limit. pcm0: chn_init(pcm0:play:0) failed: err = 19
> > > pcm0: pcm_chn_create(aggch, 1, 0xc223e700) failed
> > > pcm0: offset 0xfffa1000 exceeds limit. pcm0: chn_init(pcm0:play:0) failed: err = 19
> > > pcm0: pcm_chn_create(aggch, 1, 0xc223e700) failed
> > > pcm0: offset 0xfffa1000 exceeds limit. pcm0: chn_init(pcm0:play:0) failed: err = 19
> > > pcm0: pcm_chn_create(aggch, 1, 0xc223e700) failed
> > > pcm0: offset 0xfffa1000 exceeds limit. pcm0: chn_init(pcm0:play:0) failed: err = 19
> > > pcm0: pcm_chn_create(aggch, 1, 0xc223e700) failed
> > 
> > It sure looks to me like physaddr is getting set to zero in
> > aggch_init().  It also looks like the device_printf() format string is
> > missing a newline.
> > 
> > Try adding a device_printf() to aggch_init() to print out both physaddr
> > and ess->baseaddr so that we can figure out where the calculation
> > 	 ch->offset = physaddr - ess->baseaddr;
> > is going wrong.  The source file is /usr/src/sys/dev/sound/pci/maestro.c
> 
> This is caused by physaddr which is smaller than ess->baseaddr.
> It will often occur when memory is fragmented, I think.
> 
> Michael, would you try the newer driver?
> 	http://www.tackymt.homeip.net/~taku/maestro-latest/

Ok, I compiled and loaded your new driver, and aside from noise when
compiling, it seems to be working ok.  As an added bonus, I now have one
record channel where there was none before.  Whoo-hoo!

The aforementioned noise from compilation is attached (note that I am
using "-O2 -pipe" in /etc/make.conf - I don't know if that makes a
difference).

-- 
Mike
perl -e 'print unpack("u","88V]N=&%C=\"!I;F9O(&EN(&AE861E<G,*");'