As promised: Soft volume mixer controller

[Ariff Abdullah]

On Wed, 14 Sep 2005 12:41:20 +0200
Michael Nottebrock <lofi at freebsd.org> wrote:
> On Wednesday, 14. September 2005 04:12, Ariff Abdullah wrote:
> 
> > Can you revisit sys/dev/sound/pcm/ac97.c, look for
> > ac97_fix_volume(codec); , and comment or remove it? I just want to
> > make sure it is the fault of feeder volume, not something else.
> 
> Yes, that makes vchans work again (and the pcm mixer is binary again,
> so I  guess I commented out the right line, too).
> 
damn that feeder volume :)

> However, it seems I also found another bug related to vchans somewhere
> else in  that patchset - and one in the current unpatched 5-STABLE:
> 
> I have hw.snd.pcm0.vchans=4 set in /etc/sysctl.conf
> 
> With an unpatched kernel (5-STABLE), this gives me:
> 
> [lofi at kiste]:0:~ > ls -l /dev/dsp*
> crw-rw-rw-  1 root  wheel   30,   3 14 Sep 12:31 /dev/dsp0.0
> crw-rw-rw-  1 root  wheel   30, 0x00010003 14 Sep 12:31 /dev/dsp0.1
> crw-rw-rw-  1 root  wheel   30, 0x00020003 14 Sep 12:31 /dev/dsp0.2
> crw-rw-rw-  1 root  wheel   30, 0x00030003 14 Sep 12:31 /dev/dsp0.3
> crw-rw-rw-  1 root  wheel   30, 0x00040003 14 Sep 12:31 /dev/dsp0.4
> crw-rw-rw-  1 root  wheel   30, 0x00050003 14 Sep 12:31 /dev/dsp0.5
> crw-rw-rw-  1 root  wheel   30,   5 14 Sep 12:29 /dev/dspW0.0
> crw-rw-rw-  1 root  wheel   30, 0x00010005 14 Sep 12:29 /dev/dspW0.1
> crw-rw-rw-  1 root  wheel   30, 0x00020005 14 Sep 12:29 /dev/dspW0.2
> crw-rw-rw-  1 root  wheel   30, 0x00030005 14 Sep 12:29 /dev/dspW0.3
> crw-rw-rw-  1 root  wheel   30, 0x00040005 14 Sep 12:29 /dev/dspW0.4
> crw-rw-rw-  1 root  wheel   30, 0x00050005 14 Sep 12:29 /dev/dspW0.5
> crw-rw-rw-  1 root  wheel   30, 0x0001000b 14 Sep 12:29 /dev/dspr0.1
> 
> i.e. six virtual devices rather then four. I have no idea if that has
> always  been the case or not - I never bothered to actually cound the
> devices before,  after the four I requested were there! :-)
> 
It has always been the case. It's not a bug (depend on bug definition):
dsp0.0 = hardware channel
dsp0.1 = master / parent channel (feeder_mix_s16)
dsp0.2,3,4,5 = virtual channel

Device naming / numbering is mega confusing, I'll address this in
future.

> With a patched kernel however I get:
> 
> [lofi at kiste]:0:~ > ls -l /dev/dsp*
> crw-rw-rw-  1 root  wheel   30,   3 14 Sep 12:36 /dev/dsp0.0
> crw-rw-rw-  1 root  wheel   30, 0x00020003 14 Sep 12:36 /dev/dsp0.2
> crw-rw-rw-  1 root  wheel   30, 0x00030003 14 Sep 12:37 /dev/dsp0.3
> crw-rw-rw-  1 root  wheel   30, 0x00040003 14 Sep 12:36 /dev/dsp0.4
> crw-rw-rw-  1 root  wheel   30, 0x00050003 14 Sep 12:36 /dev/dsp0.5
> crw-rw-rw-  1 root  wheel   30, 0x00060003 14 Sep 12:36 /dev/dsp0.6
> crw-rw-rw-  1 root  wheel   30,   5 14 Sep 12:36 /dev/dspW0.0
> crw-rw-rw-  1 root  wheel   30, 0x00020005 14 Sep 12:36 /dev/dspW0.2
> crw-rw-rw-  1 root  wheel   30, 0x00030005 14 Sep 12:36 /dev/dspW0.3
> crw-rw-rw-  1 root  wheel   30, 0x00040005 14 Sep 12:36 /dev/dspW0.4
> crw-rw-rw-  1 root  wheel   30, 0x00050005 14 Sep 12:36 /dev/dspW0.5
> crw-rw-rw-  1 root  wheel   30, 0x00060005 14 Sep 12:36 /dev/dspW0.6
> crw-rw-rw-  1 root  wheel   30, 0x0002000b 14 Sep 12:36 /dev/dspr0.2
> 
> Note how /dev/dsp(W|r)0.1 are missing and instead I get /dev/dsp(W)0.6
> 
> and /dev/dspr0.2 ...
>
Same as above, only that the patched version increase the
vchan robustness at the cost of more confusion :)

Ok, back to our real problem. Can you uncomment / restore back
ac97_fix_volume(codec); , and apply this:

--- channel.c.diff BEGIN ---
--- sys/dev/sound/pcm/channel.c.orig	Wed Sep 14 19:47:43 2005
+++ sys/dev/sound/pcm/channel.c	Wed Sep 14 19:48:04 2005
@@ -884,6 +884,7 @@
 		chn_lockdestroy(c);
 
 		return ret;
+#if 0
 	} else if (c->direction == PCMDIR_PLAY &&
 			c->parentsnddev && (c->parentsnddev->flags &
SD_F_SOFTVOL) &&
 			c->parentsnddev->mixer_dev) {
@@ -899,6 +900,7 @@
 		CHN_LOCK(c);
 		chn_setvolume(c, vol & 0x7f, (vol >> 8) & 0x7f);
 		CHN_UNLOCK(c);
+#endif
 	}
 
 	return 0;
@@ -1331,7 +1333,8 @@
 	c->feederflags &= ~(1 << FEEDER_VOLUME);
 	if (c->direction == PCMDIR_PLAY &&
 			!(c->flags & (CHN_F_VIRTUAL|CHN_F_HAS_SLAVE)) &&
-			c->parentsnddev && (c->parentsnddev->flags & SD_F_SOFTVOL))
+			c->parentsnddev && (c->parentsnddev->flags & SD_F_SOFTVOL) &&
+			c->parentsnddev->mixer_dev)
 		c->feederflags |= 1 << FEEDER_VOLUME;
 	flags = c->feederflags;
 	fmtlist = chn_getcaps(c)->fmtlist;
@@ -1396,6 +1399,17 @@
 	}
 
 	sndbuf_setfmt(c->bufhard, hwfmt);
+
+	if ((flags & (1 << FEEDER_VOLUME))) {
+		int vol = 100 | (100 << 8);
+
+		CHN_UNLOCK(c);
+		if (mixer_ioctl(c->parentsnddev->mixer_dev,
+				MIXER_READ(SOUND_MIXER_PCM), (caddr_t)&vol, -1, NULL) !=
0)
+			device_printf(c->dev, "Soft Volume: Failed to read default
value\n");
+		CHN_LOCK(c);
+		chn_setvolume(c, vol & 0x7f, (vol >> 8) & 0x7f);
+	}
 
 	return 0;
 }

--- channel.c.diff END ---

--

Ariff Abdullah
MyBSD

http://www.MyBSD.org.my (IPv6/IPv4)
http://staff.MyBSD.org.my (IPv6/IPv4)
http://tomoyo.MyBSD.org.my (IPv6/IPv4)