Red Hat Bugzilla – Bug 942
Sndcnfig: libnewt dependency, CMI8330 status
Last modified: 2008-05-01 11:37:49 EDT
1) sndconfig needs (at least when I downloaded it last
weekend) a new version of libnewt (0.40). (I'll
be honest; this COULD have been fixed since, but I
haven't looked). That version of libnewt won't run
because of an unsatisfied reference to __bzero.
(With the underbars!).
I suspect that it's an unintended dependency due to
a new compiler or new option: I downloaded the sources
and compiles up and runs just fine.
When it doesn't run, it's gotten far enough to put the
tty into raw mode. (Easy enough to fix, but...)
2) The new sndconfig does recognize the CMI8830 and its
clone the HT1869V+ (which is what's on my motherboard).
However, due to strangenesses in my configuration,
the defaults don't work (they don't work on DOS,
either). (It's an IRQ conflict with an IRQ level
that no hardware claims: maybe a motherboard problem.)
isapnp DOES detect the conflict and reports it when
run by hand. However, the output from isapnp is not
reported by sndconfig. In this case, it gets far
enough that the test of playing a .wav file works,
but (e.g.) synaesthesia doen't work because of
interrupt problems (which the soundblaster module
So... (1) sndconfig should report isapnp's errors.
(2) it should report modprobe's errors.
(It appears that things are such that sndprobe makes
things sort-of work in the same boot lifetime that it
is run in, but until everything is right, sound doesn't
work after a reboot. --- Now that I've gotten the
interrupts untangled everything works fine for me.)
3) The soundblaster driver works fine with the chip, but
the setup seems strange: I was never able to get it
to work at IO 0x220, which is the default for the
DOS drivers for this chip (as far as I can tell from the
vague statements in the associated documentation).
I have it set up at 0x530 and that's fine. (There
may be a luck factor here.) (The chip as both SB and
WSS ports, and the DOS setup "wants" the WSS at 0x530
and the SB at 0x220).
4) The chip supports S/PDIF, but I was unable to find
anywhere any drivers that support it. Since the only
channel from the CD to the sound board is via the
S/PDIF channel, I needed it. I ended up hacking
dspcntrl (from the web) to do the job, but consider this
an enhancement request to the sound drivers and mixers
to support controlling that bit. (Ditto the loudness
and 3D bits.)
I'm submitting a separate bug w.r.t. isapnp at boot time,
that's a bit tangled with this.
1) is due to glibc2.1.
2) sndconfig does print isapnp errors, but only if isapnp
exits non-zero. If it exits with a status of zero, it assumes
3) Using 0x530 for the sb driver shouldn't work (it doesn't
on the CMI8330 we have here), as that's the
WSS emulation. You could try using 0x240.
4) is a driver issue; we don't maintain the drivers, per se.
Thank you for the suggestions. We have noted them and will work on
getting them into the next release.