Skip to contentSkip to conversation

Allow ioplugins to override snd_pcm_delay()

PATCH In review Pierre Ossman 3 participants 7 messages May 26, 2006

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 Allow ioplugins to override snd_pcm_delay()

Some io plug-ins might want to adjust the reported delay value and not strictly follow the current buffer usage (that's why we have two calls after all).

Allow them to specify a delay() callback and use the previous behaviour if they don't.

? ioplug-delay.patch
Index: include/pcm_ioplug.h
===================================================================
RCS file: /cvsroot/alsa/alsa-lib/include/pcm_ioplug.h,v retrieving revision 1.5

pcm_ioplug.h +4 -0
diff -u -r1.5 pcm_ioplug.h
@@ -130,6 +130,10 @@
 	 */
 	snd_pcm_sframes_t (*pointer)(snd_pcm_ioplug_t *io);
 	/**
+	 * get the delay for the running PCM; optional
+	 */
+	int (*delay)(snd_pcm_ioplug_t *io, snd_pcm_sframes_t *delayp);
+	/**
 	 * transfer the data; optional
 	 */
 	snd_pcm_sframes_t (*transfer)(snd_pcm_ioplug_t *io,

Index: src/pcm/pcm_ioplug.c
===================================================================
RCS file: /cvsroot/alsa/alsa-lib/src/pcm/pcm_ioplug.c,v retrieving revision 1.10

pcm/pcm_ioplug.c +8 -2
diff -u -r1.10 pcm_ioplug.c
@@ -106,8 +106,14 @@
 
 static int snd_pcm_ioplug_delay(snd_pcm_t *pcm, snd_pcm_sframes_t *delayp)
 {
-	snd_pcm_ioplug_hw_ptr_update(pcm);
-	*delayp = snd_pcm_mmap_hw_avail(pcm); 
+	ioplug_priv_t *io = pcm->private_data;
+
+	if (io->data->callback->delay)
+		return io->data->callback->delay(io->data, delayp);
+	else {
+		snd_pcm_ioplug_hw_ptr_update(pcm);
+		*delayp = snd_pcm_mmap_hw_avail(pcm);
+	}
 	return 0;
 }
 

_______________________________________________
Alsa-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/alsa-devel

#1 lore raw

At Fri, 26 May 2006 11:38:11 +0200,

Quoted 19 lines · Pierre Ossman

Some io plug-ins might want to adjust the reported delay value and not strictly follow the current buffer usage (that's why we have two calls after all).

Allow them to specify a delay() callback and use the previous behaviour if they don't.

? ioplug-delay.patch
Index: include/pcm_ioplug.h
===================================================================
RCS file: /cvsroot/alsa/alsa-lib/include/pcm_ioplug.h,v retrieving revision 1.5
diff -u -r1.5 pcm_ioplug.h

pcm_ioplug.h line 139

Quoted 10 lines
pcm_ioplug.h +4 -0
@@ -130,6 +130,10 @@
 	 */
 	snd_pcm_sframes_t (*pointer)(snd_pcm_ioplug_t *io);
 	/**
+	 * get the delay for the running PCM; optional
+	 */
+	int (*delay)(snd_pcm_ioplug_t *io, snd_pcm_sframes_t *delayp);
+	/**
 	 * transfer the data; optional
 	 */
 	snd_pcm_sframes_t (*transfer)(snd_pcm_ioplug_t *io,

The change is OK, but please remind that the API has been already public, so we need to keep the compatibility as much as possible. This addition will break the binary compatibility.

First, we need the increase of version number (SND_PCM_IOPLUG_VERSION) if we modify API. If the version of the object is older, either reject or allow it as compatible.

One way to keep it binary-compatible is to append the new method to the end of the callback table, and check the version before accessing the new method (and handle as NULL for older version).

Could you rewrite the patch in that manner?

Thanks,

Takashi

#2 lore raw
Quoted 9 lines · Takashi Iwai

The change is OK, but please remind that the API has been already public, so we need to keep the compatibility as much as possible. This addition will break the binary compatibility.

First, we need the increase of version number (SND_PCM_IOPLUG_VERSION) if we modify API. If the version of the object is older, either reject or allow it as compatible.

Quoted 6 lines

One way to keep it binary-compatible is to append the new method to the end of the callback table, and check the version before accessing the new method (and handle as NULL for older version).

Could you rewrite the patch in that manner?

No problem. I'll have it fixed later this afternoon.

Signature · 3 lines
Rgds
Pierre Ossman                Telephone: +46-13-21 46 00
Cendio AB                    Web: http://www.cendio.com

_______________________________________________
Alsa-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/alsa-devel

#3 lore raw
In reply to Takashi Iwai
Quoted 5 lines · Takashi Iwai

First, we need the increase of version number (SND_PCM_IOPLUG_VERSION) if we modify API. If the version of the object is older, either reject or allow it as compatible.

The headers do not detail how versions are managed here. Which component of the version number should I increase?

Signature · 3 lines
Rgds
Pierre Ossman                Telephone: +46-13-21 46 00
Cendio AB                    Web: http://www.cendio.com

_______________________________________________
Alsa-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/alsa-devel

#4 lore raw

At Fri, 26 May 2006 14:35:20 +0200,

Quoted 10 lines · Pierre Ossman
Quoted 5 lines · Takashi Iwai

First, we need the increase of version number (SND_PCM_IOPLUG_VERSION) if we modify API. If the version of the object is older, either reject or allow it as compatible.

The headers do not detail how versions are managed here. Which component of the version number should I increase?

I think TINY should suffice in this case.
It's just a minor addition.

Takashi

#5 lore raw
In reply to Pierre Ossman
Quoted 9 lines · Pierre Ossman

Some io plug-ins might want to adjust the reported delay value and not strictly follow the current buffer usage (that's why we have two calls after all).

Allow them to specify a delay() callback and use the previous behaviour if they don't.

Can you give an example of a plugin that would want to adjust the delay value?

James

#6 lore raw
Quoted 5 lines · James Courtier-Dutton
Quoted 4 lines · Pierre Ossman

Some io plug-ins might want to adjust the reported delay value and not strictly follow the current buffer usage (that's why we have two calls after all).

Quoted 8 lines
Quoted 5 lines

Allow them to specify a delay() callback and use the previous behaviour if they don't.

Can you give an example of a plugin that would want to adjust the delay value?

The polypaudio backend, which is why I submitted this patch.

The recently released padsp (esddsp-equivalent) has much better latency reporting than the ALSA to polypaudio plug-in. And we just can't have OSS applications outperforming ALSA ones. ;)

Signature · 2 lines
Pierre Ossman                Telephone: +46-13-21 46 00
Cendio AB                    Web: http://www.cendio.com

_______________________________________________
Alsa-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/alsa-devel