Firstly, the cross-post to -hackers seems appropriate but
my apologies if it isn't.
I am now certain that there is a bug in ioctl() (at least for
setting the mixer).
This started out as an attempt to fix a bug in xmms, but making a
debug version of mixer(8) showed it to be affected the same way. The
following illustrates the problem, notice that the pcm is nearly
always set to a value different to that which is passed on the
command-line.
I haven't submitted a PR for this as I'm not 100% certain how for
something like this.
% mixer pcm
Mixer pcm is currently set to 91:91
% mixer pcm 90:90
Setting the mixer pcm to 90:90.
% mixer pcm
Mixer pcm is currently set to 91:91
% mixer pcm 88:88
Setting the mixer pcm to 88:88.
% mixer pcm
Mixer pcm is currently set to 88:88
% mixer pcm 86:86
Setting the mixer pcm to 86:86.
% mixer pcm
Mixer pcm is currently set to 88:88
% mixer pcm 84:84
Setting the mixer pcm to 84:84.
% mixer pcm
Mixer pcm is currently set to 85:85
%
On Sat, Feb 19, 2000 at 02:55:18PM +0000, Mark Ovens wrote:
> # uname -a
> FreeBSD marder-1 3.4-STABLE FreeBSD 3.4-STABLE #0:
> Fri Feb 18 18:14:43 GMT 2000
> root@marder-1:/usr/src/sys/compile/MARDER-1 i386
>
>
> I'm trying to debug a problem in the xmms port whereby the volume
> doesn't get set correctly.
>
> It goes wrong after the ioctl() call to set the volume in /dev/mixer.
>
> >From /usr/include/machine/soundcard.h:
>
> #define SOUND_MIXER_PCM 4
> #define MIXER_WRITE(dev) _IOWR('M', dev, int)
> #define SOUND_MIXER_WRITE_PCM MIXER_WRITE(SOUND_MIXER_PCM)
>
> and the code in question is
>
> int fd, v, l, r;
>
> fd = open(devname, O_RDONLY);
>
> l = 90; r = 90;
>
> v = (r << 8) | l; /* v == 23130 */
>
> ioctl(fd, SOUND_MIXER_WRITE_PCM, &v);
>
> /* now v == 23387, which is (91 << 8) | 91 */
>
> Before the ioctl() call the output from mixer(8) was:
>
> Mixer pcm is currently set to 88:88
>
> and after:
>
> Mixer pcm is currently set to 91:91
>
> when it should be 90:90
>
> The appears to be a bug in ioctl(), but is it?
>
> /dev/mixer is being opened read-only but we appear to be writing to it
> OK, albeit the wrong values.
>
> I'm getting a bit out of my depth here so a bit of help would be
> appreciated. Is this a bug in ioctl()? If not, any suggestions as to
> what I should do now to find the cause of the problem?
>
> Thanks.
>
> --
> Microsoft: Where do you want to go today?
> Linux: Where do you want to go tomorrow?
> BSD: Are you guys coming, or what?
> -Poster at LinuxWorld 2000
> ________________________________________________________________
> FreeBSD - The Power To Serve http://www.freebsd.org
> My Webpage http://ukug.uk.freebsd.org/~mark/
> mailto:[EMAIL PROTECTED] http://www.radan.com
>
>
>
> To Unsubscribe: send mail to [EMAIL PROTECTED]
> with "unsubscribe freebsd-stable" in the body of the message
--
Microsoft: Where do you want to go today?
Linux: Where do you want to go tomorrow?
BSD: Are you guys coming, or what?
-Poster at LinuxWorld 2000
________________________________________________________________
FreeBSD - The Power To Serve http://www.freebsd.org
My Webpage http://ukug.uk.freebsd.org/~mark/
mailto:[EMAIL PROTECTED] http://www.radan.com
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message
