Skip to contentSkip to conversation

amidi delay option

PATCH Superseded Felipe F. Tonello 4 participants 14 messages Aug 15, 2016 – Aug 23, 2016

v5 of 6 Superseded Latest is v6 →

lore
Share & tools
lore
#0 lore raw

v5 cover amidi delay option

Small improvement to amidi, adding useful functionality.

changes from v4:
* simplified code

changes from v3:
* fixed a bug if memchr() returns NULL

changes from v2:
* only check of end of sysex (0xf7)
* use memchr() instead of direct loop approach
* fixed commit message

changes from v1:
* remove delay in between any MIDI message, only SysEx
* added patch to use getopt_long() arguments

Felipe F. Tonello (2):
amidi: add delay option
amidi: use GNU getopt_long() proper argument names

 amidi/amidi.1 | 14 ++++++++++++++
 amidi/amidi.c | 61 +++++++++++++++++++++++++++++++++++++++++++----------------
 2 files changed, 59 insertions(+), 16 deletions(-)
Signature · 1 line
2.9.3
#1 lore raw

PATCH v5 1/2 amidi: add delay option

This patch adds a new options to amidi tool to add a delay (in milliseconds) to each SysEx message.

This is very useful when sending firmware updates to a remote device via SysEx or any other use that requires this delay in between messages.

---
 amidi/amidi.1 | 14 ++++++++++++++
 amidi/amidi.c | 39 ++++++++++++++++++++++++++++++++++-----
 2 files changed, 48 insertions(+), 5 deletions(-)
amidi/amidi.1 +14 -0
index 1b4cfb1ea956..391008177610 100644
@@ -111,6 +111,12 @@ to stop receiving data.
 Does not ignore Active Sensing bytes (FEh) when saving or printing
 received MIDI commands.
 
+.TP
+.I \-D, \-\-delay=mseconds
+Adds a delay in between each SysEx message sent to a device. It is
+useful when sending firmware updates via SysEx messages to a remote
+device.
+
 .SH EXAMPLES
 
 .TP
@@ -121,6 +127,14 @@ to port
 .I hw:0.
 
 .TP
+.B amidi \-p hw:1,0,0 -s firmware.syx \-D 100
+will send the MIDI commands in
+.I firmware.syx
+to port
+.I hw:1,0,0
+with 100 milliseconds delay in between each SysEx message.
+
+.TP
 .B amidi \-S 'F0 43 10 4C 00 00 7E 00 F7'
 sends an XG Reset to the default port.
 
amidi/amidi.c +34 -5
index cedf18c5fd39..bdb04c50a50b 100644
@@ -48,6 +48,7 @@ static int receive_file;
 static int dump;
 static int timeout;
 static int stop;
+static int delay;
 static snd_rawmidi_t *input, **inputp;
 static snd_rawmidi_t *output, **outputp;
 
@@ -77,7 +78,8 @@ static void usage(void)
 		"-d, --dump             print received data as hexadecimal bytes\n"
 		"-t, --timeout=seconds  exits when no data has been received\n"
 		"                       for the specified duration\n"
-		"-a, --active-sensing   don't ignore active sensing bytes\n");
+		"-a, --active-sensing   don't ignore active sensing bytes\n"
+		"-D, --delay=mseconds   delay in between each SysEx message\n");
 }
 
 static void version(void)
@@ -406,7 +408,7 @@ static void add_send_hex_data(const char *str)
 
 int main(int argc, char *argv[])
 {
-	static const char short_options[] = "hVlLp:s:r:S::dt:a";
+	static const char short_options[] = "hVlLp:s:r:S::dt:aD:";
 	static const struct option long_options[] = {
 		{"help", 0, NULL, 'h'},
 		{"version", 0, NULL, 'V'},
@@ -419,6 +421,7 @@ int main(int argc, char *argv[])
 		{"dump", 0, NULL, 'd'},
 		{"timeout", 1, NULL, 't'},
 		{"active-sensing", 0, NULL, 'a'},
+		{"delay", 1, NULL, 'D'},
 		{ }
 	};
 	int c, err, ok = 0;
@@ -463,6 +466,9 @@ int main(int argc, char *argv[])
 		case 'a':
 			ignore_active_sensing = 0;
 			break;
+		case 'D':
+			delay = atoi(optarg);
+			break;
 		default:
 			error("Try `amidi --help' for more information.");
 			return 1;
@@ -538,9 +544,32 @@ int main(int argc, char *argv[])
 			error("cannot set blocking mode: %s", snd_strerror(err));
 			goto _exit;
 		}
-		if ((err = snd_rawmidi_write(output, send_data, send_data_length)) < 0) {
-			error("cannot send data: %s", snd_strerror(err));
-			goto _exit;
+		if (!delay) {
+			if ((err = snd_rawmidi_write(output, send_data, send_data_length)) < 0) {
+				error("cannot send data: %s", snd_strerror(err));
+				return err;
+			}
+		} else {
+			char *data = send_data;
+
+			while (data < (send_data + send_data_length)) {
+				int len = send_data + send_data_length - data;
+				char *temp;
+
+				if (data > send_data)
+					usleep(delay * 1000);
+
+				/* find end of SysEx */
+				if ((temp = memchr(data, 0xf7, len)))
+					len = temp - data + 1;
+
+				if ((err = snd_rawmidi_write(output, data, len)) < 0) {
+					error("cannot send data: %s", snd_strerror(err));
+					return err;
+				}
+
+				data += len;
+			}
 		}
 	}
 
Signature · 1 line
2.9.3
#2 lore raw

Re: [PATCH v5 1/2] amidi: add delay option

amidi/amidi.c int main(int argc, char *argv[]) · line 560

Quoted 9 lines · Felipe F. Tonello

+.I \-D, \-\-delay=mseconds

amidi/amidi.c +7 -0
+Adds a delay in between each SysEx message sent to a device.
 
+			while (data < (send_data + send_data_length)) {
+				int len = send_data + send_data_length - data;
+				char *temp;
+
+				if (data > send_data)
+					usleep(delay * 1000);

The MIDI data is sent asynchonously, so this delay is not _between_ messages. You'd have to add 320 µs for any byte not yet sent (I guess you have to use snd_rawmidi_status_get_avail() to find out how many have been sent).

amidi/amidi.c int main(int argc, char *argv[]) · line 563

Quoted 2 lines
amidi/amidi.c +2 -0
+				/* find end of SysEx */
+				if ((temp = memchr(data, 0xf7, len)))

This would be clearer with "!= NULL".

amidi/amidi.c int main(int argc, char *argv[]) · line 574

Quoted 11 lines
amidi/amidi.c +9 -0
+					len = temp - data + 1;
+
+				if ((err = snd_rawmidi_write(output, data, len)) < 0) {
+					error("cannot send data: %s", snd_strerror(err));
+					return err;
+				}
+
+				data += len;
+			}
 		}
 	}

Regards,
Clemens

#3 lore raw

Re: [PATCH v5 1/2] amidi: add delay option

Hi Clemens,

Quoted 4 lines · Clemens Ladisch
Quoted 2 lines · Felipe F. Tonello

+.I \-D, \-\-delay=mseconds
+Adds a delay in between each SysEx message sent to a device.

Quoted 10 lines

amidi/amidi.c int main(int argc, char *argv[]) · line 560

Quoted 6 lines
amidi/amidi.c +6 -0
+			while (data < (send_data + send_data_length)) {
+				int len = send_data + send_data_length - data;
+				char *temp;
+
+				if (data > send_data)
+					usleep(delay * 1000);

The MIDI data is sent asynchonously, so this delay is not _between_ messages. You'd have to add 320 µs for any byte not yet sent (I guess you have to use snd_rawmidi_status_get_avail() to find out how many have been sent).

Right. Or I can just use snd_rawmidi_drain()? Otherwise it will add too much complexity.

Quoted 6 lines

amidi/amidi.c int main(int argc, char *argv[]) · line 563

Quoted 4 lines
amidi/amidi.c +2 -0
+				/* find end of SysEx */
+				if ((temp = memchr(data, 0xf7, len)))

This would be clearer with "!= NULL".

Ok.

Quoted 13 lines

amidi/amidi.c int main(int argc, char *argv[]) · line 574

Quoted 12 lines
amidi/amidi.c +9 -0
+					len = temp - data + 1;
+
+				if ((err = snd_rawmidi_write(output, data, len)) < 0) {
+					error("cannot send data: %s", snd_strerror(err));
+					return err;
+				}
+
+				data += len;
+			}
 		}
 	}

Thanks

Signature · 1 line
Felipe
#4 lore raw

Re: [PATCH v5 1/2] amidi: add delay option

Quoted 10 lines · Felipe Ferreri Tonello
Quoted 7 lines · Clemens Ladisch

amidi/amidi.c int main(int argc, char *argv[]) · line 560

Quoted 2 lines · Felipe F. Tonello
amidi/amidi.c +2 -0
+				if (data > send_data)
+					usleep(delay * 1000);

The MIDI data is sent asynchonously, so this delay is not _between_ messages. You'd have to add 320 µs for any byte not yet sent (I guess you have to use snd_rawmidi_status_get_avail() to find out how many have been sent).

Right. Or I can just use snd_rawmidi_drain()?

This function not only drains the ALSA buffer, but also waits for the hardware FIFO to empty. And most drivers do not implement this, in which case the framework just does a fixed wait of 50 ms, which would correspond to a FIFO size of 156 bytes.

Regards,
Clemens

#5 lore raw

Re: [PATCH v5 1/2] amidi: add delay option

Hi Clemens,

Quoted 10 lines · Clemens Ladisch
Quoted 9 lines · Felipe Ferreri Tonello
Quoted 7 lines · Clemens Ladisch

amidi/amidi.c int main(int argc, char *argv[]) · line 560

Quoted 2 lines · Felipe F. Tonello
amidi/amidi.c +2 -0
+				if (data > send_data)
+					usleep(delay * 1000);

The MIDI data is sent asynchonously, so this delay is not _between_ messages. You'd have to add 320 µs for any byte not yet sent (I guess you have to use snd_rawmidi_status_get_avail() to find out how many have been sent).

Quoted 7 lines
Quoted 1 line

Right. Or I can just use snd_rawmidi_drain()?

This function not only drains the ALSA buffer, but also waits for the hardware FIFO to empty. And most drivers do not implement this, in which case the framework just does a fixed wait of 50 ms, which would correspond to a FIFO size of 156 bytes.

Ok, so does that mean that it is not safe to use drain()? If that is the case, than I will use snd_rawmidi_status_get_avail().

Signature · 1 line
Felipe
#6 lore raw

Re: [PATCH v5 1/2] amidi: add delay option

Quoted 11 lines · Felipe Ferreri Tonello
Quoted 10 lines · Clemens Ladisch
Quoted 9 lines · Felipe Ferreri Tonello
Quoted 7 lines · Clemens Ladisch

amidi/amidi.c int main(int argc, char *argv[]) · line 560

Quoted 2 lines · Felipe F. Tonello
amidi/amidi.c +2 -0
+				if (data > send_data)
+					usleep(delay * 1000);

The MIDI data is sent asynchonously, so this delay is not _between_ messages. You'd have to add 320 µs for any byte not yet sent (I guess you have to use snd_rawmidi_status_get_avail() to find out how many have been sent).

Quoted 8 lines
Quoted 6 lines
Quoted 1 line

Right. Or I can just use snd_rawmidi_drain()?

This function not only drains the ALSA buffer, but also waits for the hardware FIFO to empty. And most drivers do not implement this, in which case the framework just does a fixed wait of 50 ms, which would correspond to a FIFO size of 156 bytes.

Ok, so does that mean that it is not safe to use drain()?

That additional 50 ms delay might be safe, but not what you want.

Regards,
Clemens

#7 lore raw

Re: [PATCH v5 1/2] amidi: add delay option

Hi Clemens,

Quoted 14 lines · Clemens Ladisch
Quoted 13 lines · Felipe Ferreri Tonello
Quoted 12 lines · Clemens Ladisch
Quoted 10 lines · Felipe Ferreri Tonello
Quoted 7 lines · Clemens Ladisch

amidi/amidi.c int main(int argc, char *argv[]) · line 560

Quoted 2 lines · Felipe F. Tonello
amidi/amidi.c +2 -0
+				if (data > send_data)
+					usleep(delay * 1000);

The MIDI data is sent asynchonously, so this delay is not _between_ messages. You'd have to add 320 µs for any byte not yet sent (I guess you have to use snd_rawmidi_status_get_avail() to find out how many have been sent).

Right. Or I can just use snd_rawmidi_drain()?

Quoted 8 lines
Quoted 6 lines
Quoted 4 lines

This function not only drains the ALSA buffer, but also waits for the hardware FIFO to empty. And most drivers do not implement this, in which case the framework just does a fixed wait of 50 ms, which would correspond to a FIFO size of 156 bytes.

Ok, so does that mean that it is not safe to use drain()?

That additional 50 ms delay might be safe, but not what you want.

Ok.

BTW: Why 320 µs for any byte not yet sent?

Thanks,

Signature · 1 line
Felipe
#8 lore raw

Re: [PATCH v5 1/2] amidi: add delay option

Quoted 1 line · Felipe Ferreri Tonello

Why 320 µs for any byte not yet sent?

31250 baud; 10 bits per byte.

Regards,
Clemens

#9 lore raw

Re: [PATCH v5 1/2] amidi: add delay option

Hi Clemens,

Quoted 4 lines · Clemens Ladisch
Quoted 1 line · Felipe Ferreri Tonello

Why 320 µs for any byte not yet sent?

31250 baud; 10 bits per byte.

Right. Thanks!

Signature · 1 line
Felipe
#10 lore raw

PATCH v5 2/2 amidi: use GNU getopt_long() proper argument names

In reply to Felipe F. Tonello

This has no affect besides more clarity to the code.

---
 amidi/amidi.c | 24 ++++++++++++------------
 1 file changed, 12 insertions(+), 12 deletions(-)
amidi/amidi.c +12 -12
index bdb04c50a50b..0be5e63119c0 100644
@@ -410,18 +410,18 @@ int main(int argc, char *argv[])
 {
 	static const char short_options[] = "hVlLp:s:r:S::dt:aD:";
 	static const struct option long_options[] = {
-		{"help", 0, NULL, 'h'},
-		{"version", 0, NULL, 'V'},
-		{"list-devices", 0, NULL, 'l'},
-		{"list-rawmidis", 0, NULL, 'L'},
-		{"port", 1, NULL, 'p'},
-		{"send", 1, NULL, 's'},
-		{"receive", 1, NULL, 'r'},
-		{"send-hex", 2, NULL, 'S'},
-		{"dump", 0, NULL, 'd'},
-		{"timeout", 1, NULL, 't'},
-		{"active-sensing", 0, NULL, 'a'},
-		{"delay", 1, NULL, 'D'},
+		{"help", no_argument, NULL, 'h'},
+		{"version", no_argument, NULL, 'V'},
+		{"list-devices", no_argument, NULL, 'l'},
+		{"list-rawmidis", no_argument, NULL, 'L'},
+		{"port", required_argument, NULL, 'p'},
+		{"send", required_argument, NULL, 's'},
+		{"receive", required_argument, NULL, 'r'},
+		{"send-hex", optional_argument, NULL, 'S'},
+		{"dump", no_argument, NULL, 'd'},
+		{"timeout", required_argument, NULL, 't'},
+		{"active-sensing", no_argument, NULL, 'a'},
+		{"delay", required_argument, NULL, 'D'},
 		{ }
 	};
 	int c, err, ok = 0;
Signature · 1 line
2.9.3
#11 lore raw
In reply to Felipe F. Tonello

Hi,

Quoted 17 lines · Felipe F. Tonello

Small improvement to amidi, adding useful functionality.

changes from v4:
* simplified code

changes from v3:
* fixed a bug if memchr() returns NULL

changes from v2:
* only check of end of sysex (0xf7)
* use memchr() instead of direct loop approach
* fixed commit message

changes from v1:
* remove delay in between any MIDI message, only SysEx
* added patch to use getopt_long() arguments

Quoted 7 lines

Felipe F. Tonello (2):
amidi: add delay option
amidi: use GNU getopt_long() proper argument names

 amidi/amidi.1 | 14 ++++++++++++++
 amidi/amidi.c | 61 +++++++++++++++++++++++++++++++++++++++++++----------------
 2 files changed, 59 insertions(+), 16 deletions(-)

Please notice the other developers' work. Clemens already posted his patchset with similar improvements.

http://mailman.alsa-project.org/pipermail/alsa-devel/2016-August/111737.html

And I reviewed and tested it.
http://mailman.alsa-project.org/pipermail/alsa-devel/2016-August/111745.html

It's not merged yet, because subsystem maintainer has a summer vacation.
http://mailman.alsa-project.org/pipermail/alsa-devel/2016-August/111741.html

It's better for you to wait for merging or work based on Clemens' patchset, I think.

Regards

Takashi Sakamoto

#12 lore raw
Quoted 34 lines · Takashi Sakamoto

Hi,

Quoted 17 lines · Felipe F. Tonello

Small improvement to amidi, adding useful functionality.

changes from v4:
* simplified code

changes from v3:
* fixed a bug if memchr() returns NULL

changes from v2:
* only check of end of sysex (0xf7)
* use memchr() instead of direct loop approach
* fixed commit message

changes from v1:
* remove delay in between any MIDI message, only SysEx
* added patch to use getopt_long() arguments

Quoted 7 lines

Felipe F. Tonello (2):
amidi: add delay option
amidi: use GNU getopt_long() proper argument names

 amidi/amidi.1 | 14 ++++++++++++++
 amidi/amidi.c | 61 +++++++++++++++++++++++++++++++++++++++++++----------------
 2 files changed, 59 insertions(+), 16 deletions(-)

Please notice the other developers' work. Clemens already posted his patchset with similar improvements.

http://mailman.alsa-project.org/pipermail/alsa-devel/2016-August/111737.html

Quoted 8 lines

And I reviewed and tested it.
http://mailman.alsa-project.org/pipermail/alsa-devel/2016-August/111745.html

It's not merged yet, because subsystem maintainer has a summer vacation.
http://mailman.alsa-project.org/pipermail/alsa-devel/2016-August/111741.html

It's better for you to wait for merging or work based on Clemens' patchset, I think.

Yep, now Clemens' patches were merged to git tree. Please rebase and resubmit if anything is missing.

thanks,

Takashi

#13 lore raw

Hi Takashi,

Quoted 39 lines · Takashi Iwai
Quoted 37 lines · Takashi Sakamoto

Hi,

Quoted 17 lines · Felipe F. Tonello

Small improvement to amidi, adding useful functionality.

changes from v4:
* simplified code

changes from v3:
* fixed a bug if memchr() returns NULL

changes from v2:
* only check of end of sysex (0xf7)
* use memchr() instead of direct loop approach
* fixed commit message

changes from v1:
* remove delay in between any MIDI message, only SysEx
* added patch to use getopt_long() arguments

Quoted 7 lines

Felipe F. Tonello (2):
amidi: add delay option
amidi: use GNU getopt_long() proper argument names

 amidi/amidi.1 | 14 ++++++++++++++
 amidi/amidi.c | 61 +++++++++++++++++++++++++++++++++++++++++++----------------
 2 files changed, 59 insertions(+), 16 deletions(-)

Please notice the other developers' work. Clemens already posted his patchset with similar improvements.

http://mailman.alsa-project.org/pipermail/alsa-devel/2016-August/111737.html

And I reviewed and tested it.
http://mailman.alsa-project.org/pipermail/alsa-devel/2016-August/111745.html

Quoted 8 lines
Quoted 5 lines

It's not merged yet, because subsystem maintainer has a summer vacation.
http://mailman.alsa-project.org/pipermail/alsa-devel/2016-August/111741.html

It's better for you to wait for merging or work based on Clemens' patchset, I think.

Yep, now Clemens' patches were merged to git tree. Please rebase and resubmit if anything is missing.

Ok.

Signature · 1 line
Felipe