$ ls -als /dev/sound
total 0
0 drwxr-xr-x 1 root root 0 Jan 1 1970 .
0 drwxr-xr-x 1 root root 0 Jan 1 1970 ..
0 crw-rw-rw- 1 root root 14, 14 Jan 1 1970 admmidi
0 crw-rw-rw- 1 root root 14, 12 Jan 1 1970 adsp
0 crw-rw-rw- 1 root root 14, 13 Jan 1 1970 amidi
0 crw-rw-rw- 1 root root 14, 4 Jan 1 1970 audio
0 crw-rw-rw- 1 root root 14, 20 Jan 1 1970 audio1
0 crw-rw-rw- 1 root root 14, 9 Jan 1 1970 dmmidi
0 crw-rw-rw- 1 root root 14, 3 Jan 1 1970 dsp
0 crw-rw-rw- 1 root root 14, 19 Jan 1 1970 dsp1
0 crw-rw-rw- 1 root root 14, 2 Jan 1 1970 midi
0 crw-rw-rw- 1 root root 14, 0 Jan 1 1970 mixer
0 crw-rw-rw- 1 root root 14, 16 Jan 1 1970 mixer1
0 crw-rw-rw- 1 root root 14, 1 Jan 1 1970 sequencer
0 crw-rw-rw- 1 root root 14, 8 Jan 1 1970 sequencer2
Without it, my /dev/sound directory looks like this:
$ ls -als /dev/sound
total 0
0 drwxr-xr-x 1 root root 0 Jan 1 1970 .
0 drwxr-xr-x 1 root root 0 Jan 1 1970 ..
0 crw-r--r-- 1 root root 14, 14 Jan 1 1970 admmidi
0 crw-r--r-- 1 root root 14, 12 Jan 1 1970 adsp
0 crw-r--r-- 1 root root 14, 13 Jan 1 1970 amidi
0 crw-rw-rw- 1 root root 14, 4 Jan 1 1970 audio
0 crw-r--r-- 1 root root 14, 9 Jan 1 1970 dmmidi
0 crw-rw-rw- 1 root root 14, 3 Jan 1 1970 dsp
0 crw------- 1 root root 14, 2 Jan 1 1970 midi00
0 crw-rw-rw- 1 root root 14, 0 Jan 1 1970 mixer
0 crw-rw-rw- 1 root root 14, 1 Jan 1 1970 sequencer
0 crw-r--r-- 1 root root 14, 8 Jan 1 1970 sequencer2
0 crw------- 1 root root 14, 16 Jan 1 1970 unknown
and the following errors are written into my kernel log:
Jan 6 18:10:27 twopit kernel: devfs: devfs_register(unknown): could not append to parent, err: -17
Jan 6 18:10:27 twopit kernel: devfs: devfs_register(unknown): could not append to parent, err: -17
These errors are presumably caused by devices "mixer1", "dsp1" and
"audio1" all fighting over the "unknown" slot. Could someone please
apply the patch? Any feedback on any cleanups it might need would
also be appreciated. Note that the uniqueness of the sequencer and
sequencer2 devices is preserved.
Cheers,
Chris
--- linux-2.4.17/drivers/sound/sound_core.c.orig Sun Jan 6 17:31:30 2002
+++ linux-2.4.17/drivers/sound/sound_core.c Sun Jan 6 17:39:43 2002
@@ -17,7 +17,7 @@
* plug into this. The fact they dont all go via OSS doesn't mean
* they don't have to implement the OSS API. There is a lot of logic
* to keeping much of the OSS weight out of the code in a compatibility
- * module, but its up to the driver to rember to load it...
+ * module, but it's up to the driver to remember to load it...
*
* The code provides a set of functions for registration of devices
* by type. This is done rather than providing a single call so that
@@ -173,10 +173,10 @@
return r;
}
- if (r == low)
+ if ( r < SOUND_STEP )
sprintf (name_buf, "%s", name);
else
- sprintf (name_buf, "%s%d", name, (r - low) / SOUND_STEP);
+ sprintf (name_buf, "%s%d", name, (r / SOUND_STEP));
s->de = devfs_register (devfs_handle, name_buf,
DEVFS_FL_NONE, SOUND_MAJOR, s->unit_minor,
S_IFCHR | mode, fops, NULL);
@@ -231,17 +231,20 @@
int register_sound_special(struct file_operations *fops, int unit)
{
- char *name;
+ const int chain = (unit & 0x0F);
+ int max_unit = chain + 128;
+ const char *name;
- switch (unit) {
+ switch (chain) {
case 0:
name = "mixer";
break;
case 1:
name = "sequencer";
+ max_unit = unit + 1;
break;
case 2:
- name = "midi00";
+ name = "midi";
break;
case 3:
name = "dsp";
@@ -260,6 +263,7 @@
break;
case 8:
name = "sequencer2";
+ max_unit = unit + 1;
break;
case 9:
name = "dmmidi";
@@ -280,10 +284,10 @@
name = "admmidi";
break;
default:
- name = "unknown";
+ name = "unknownX";
break;
}
- return sound_insert_unit(&chains[unit&15], fops, -1, unit, unit+1,
+ return sound_insert_unit(&chains[chain], fops, -1, unit, max_unit,
name, S_IRUSR | S_IWUSR);
}
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/