Re: rawmidi

[Takashi Iwai <tiwai@suse.de>]

[-- Attachment #1: Type: text/plain, Size: 1445 bytes --]

At Fri, 8 Oct 2004 11:26:39 -0400 (EDT),
Patrick Dumais wrote:
> 
> I noticed that some drivers tend to send "batch" note-on commands. I mean:
> instead of sending 90h,01h,7Fh everytime, it sends a first 90,01,7f then
> just 02,7f etc... It sends the noteon command only once and assumes that
> your application will set itself in a "noteon receive" mode. This is very
> annoying because my application expects to receive a command byte each
> time. the emu10k1 driver doesn't do that, but snd-virmidi does.

Yes, virmidi reduces the command byte when the identical
running-status is kept.

Note that this is the correct MIDI implementation.  Many devices
output in this way (although many don't).


> I'm using virmidi because I want 2 apps to read my sbLive midi port. I
> aconnect this port to 2 virmidi ports then my apps read from those ports.
> 
> Is there any way that I can configure virmidi so that it sends commands
> the conventional way, or any other way that I could have 2 applications
> reading on the same midi port? (using the alsa sequencer interface is not
> an alternative method that I want to consider for the scope of my
> applications)

The attached (untested) patch will add no_status option to snd-virmidi
module.  When this is set to 1, the running-status is always output as
you like.

But, again, reducing the identical running-status is the feature of
MIDI byte stream.  Your app should really handle it.


Takashi

[-- Attachment #2: Type: text/plain, Size: 2488 bytes --]

Index: alsa-kernel/core/seq/seq_virmidi.c
===================================================================
RCS file: /suse/tiwai/cvs/alsa/alsa-kernel/core/seq/seq_virmidi.c,v
retrieving revision 1.12
diff -u -r1.12 seq_virmidi.c
--- alsa-kernel/core/seq/seq_virmidi.c	29 Jun 2004 16:01:14 -0000	1.12
+++ alsa-kernel/core/seq/seq_virmidi.c	8 Oct 2004 15:48:46 -0000
@@ -212,6 +212,8 @@
 		kfree(vmidi);
 		return -ENOMEM;
 	}
+	if (rdev->flags & SNDRV_VIRMIDI_NO_STATUS)
+		snd_midi_event_no_status(vmidi->parser, 1);
 	vmidi->seq_mode = rdev->seq_mode;
 	vmidi->client = rdev->client;
 	vmidi->port = rdev->port;	
Index: alsa-kernel/drivers/virmidi.c
===================================================================
RCS file: /suse/tiwai/cvs/alsa/alsa-kernel/drivers/virmidi.c,v
retrieving revision 1.12
diff -u -r1.12 virmidi.c
--- alsa-kernel/drivers/virmidi.c	30 Jun 2004 14:20:20 -0000	1.12
+++ alsa-kernel/drivers/virmidi.c	8 Oct 2004 15:47:03 -0000
@@ -65,6 +65,7 @@
 static char *id[SNDRV_CARDS] = SNDRV_DEFAULT_STR;	/* ID for this card */
 static int enable[SNDRV_CARDS] = {1, [1 ... (SNDRV_CARDS - 1)] = 0};
 static int midi_devs[SNDRV_CARDS] = {[0 ... (SNDRV_CARDS - 1)] = 4};
+static int no_status[SNDRV_CARDS];
 static int boot_devs;
 
 module_param_array(index, int, boot_devs, 0444);
@@ -75,6 +76,8 @@
 MODULE_PARM_DESC(enable, "Enable this soundcard.");
 module_param_array(midi_devs, int, boot_devs, 0444);
 MODULE_PARM_DESC(midi_devs, "MIDI devices # (1-8)");
+module_param_array(no_status, int, boot_devs, 0444);
+MODULE_PARM_DESC(no_status, "Add running-status always.");
 
 typedef struct snd_card_virmidi {
 	snd_card_t *card;
@@ -109,6 +112,8 @@
 		if ((err = snd_virmidi_new(card, idx, &rmidi)) < 0)
 			goto __nodev;
 		rdev = rmidi->private_data;
+		if (no_status[dev])
+			rdev->flags |= SNDRV_VIRMIDI_NO_STATUS;
 		vmidi->midi[idx] = rmidi;
 		strcpy(rmidi->name, "Virtual Raw MIDI");
 		rdev->seq_mode = SNDRV_VIRMIDI_SEQ_DISPATCH;
Index: alsa-kernel/include/seq_virmidi.h
===================================================================
RCS file: /suse/tiwai/cvs/alsa/alsa-kernel/include/seq_virmidi.h,v
retrieving revision 1.3
diff -u -r1.3 seq_virmidi.h
--- alsa-kernel/include/seq_virmidi.h	4 Jul 2002 09:41:50 -0000	1.3
+++ alsa-kernel/include/seq_virmidi.h	8 Oct 2004 15:46:42 -0000
@@ -47,6 +47,7 @@
 
 #define SNDRV_VIRMIDI_SUBSCRIBE		(1<<0)
 #define SNDRV_VIRMIDI_USE		(1<<1)
+#define SNDRV_VIRMIDI_NO_STATUS		(1<<2)
 
 /*
  * device record: