Skip to contentSkip to conversation

ASoC: soc-compress: add a config item for soc-compress

PATCH Superseded Jie Yang 5 participants 24 messages Jun 15, 2015 – Sep 15, 2015

Subsystems (4)

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore

Jie Yang

author
#0 lore raw

PATCH v2 1/2 ASoC: soc-compress: add a config item for soc-compress

We don't always need soc-compress in soc, here add a config item SND_SOC_COMPRESS, please add 'select SND_SOC_COMPRESS' to driver Kconfig when it is needed.

---
 include/sound/soc.h     | 7 +++++++
 sound/soc/Kconfig       | 5 ++++-
 sound/soc/Makefile      | 3 ++-
 sound/soc/intel/Kconfig | 1 +
 4 files changed, 14 insertions(+), 2 deletions(-)
include/sound/soc.h +7 -0
index 93df8bf..cc1cd4f 100644
@@ -440,7 +440,14 @@ int snd_soc_platform_read(struct snd_soc_platform *platform,
 int snd_soc_platform_write(struct snd_soc_platform *platform,
 					unsigned int reg, unsigned int val);
 int soc_new_pcm(struct snd_soc_pcm_runtime *rtd, int num);
+#if IS_ENABLED(CONFIG_SND_SOC_COMPRESS)
 int soc_new_compress(struct snd_soc_pcm_runtime *rtd, int num);
+#else
+static inline int soc_new_compress(struct snd_soc_pcm_runtime *rtd, int num)
+{
+       return -EPERM;
+}
+#endif
 
 struct snd_pcm_substream *snd_soc_get_dai_substream(struct snd_soc_card *card,
 		const char *dai_link, int stream);
sound/soc/Kconfig +4 -1
index e2828e1..a124759 100644
@@ -9,7 +9,6 @@ menuconfig SND_SOC
 	select SND_JACK if INPUT=y || INPUT=SND
 	select REGMAP_I2C if I2C
 	select REGMAP_SPI if SPI_MASTER
-	select SND_COMPRESS_OFFLOAD
 	---help---
 
 	  If you want ASoC support, you should say Y here and also to the
@@ -30,6 +29,10 @@ config SND_SOC_GENERIC_DMAENGINE_PCM
 	bool
 	select SND_DMAENGINE_PCM
 
+config SND_SOC_COMPRESS
+	tristate
+	select SND_COMPRESS_OFFLOAD
+
 # All the supported SoCs
 source "sound/soc/adi/Kconfig"
 source "sound/soc/atmel/Kconfig"
sound/soc/Makefile +2 -1
index a0e1ee6..184c1e6 100644
@@ -1,6 +1,7 @@
 snd-soc-core-objs := soc-core.o soc-dapm.o soc-jack.o soc-cache.o soc-utils.o
-snd-soc-core-objs += soc-pcm.o soc-compress.o soc-io.o soc-devres.o soc-ops.o
+snd-soc-core-objs += soc-pcm.o soc-io.o soc-devres.o soc-ops.o
 snd-soc-core-objs += soc-topology.o
+snd-soc-core-$(CONFIG_SND_SOC_COMPRESS) += soc-compress.o
 
 ifneq ($(CONFIG_SND_SOC_GENERIC_DMAENGINE_PCM),)
 snd-soc-core-objs += soc-generic-dmaengine-pcm.o
sound/soc/intel/Kconfig +1 -0
index 791953f..e559174 100644
@@ -12,6 +12,7 @@ config SND_MFLD_MACHINE
 
 config SND_SST_MFLD_PLATFORM
 	tristate
+	select SND_SOC_COMPRESS
 
 config SND_SST_IPC
 	tristate
Signature · 1 line
1.9.1
author
#1 lore raw

PATCH v2 2/2 ASoC: soc-compress: split soc-compress to a module

Split soc-compress into a separate module so we only compile/load it when it's needed.

---
 sound/soc/Makefile       | 4 +++-
 sound/soc/soc-compress.c | 8 ++++++++
 2 files changed, 11 insertions(+), 1 deletion(-)
sound/soc/Makefile +3 -1
index 184c1e6..2936088 100644
@@ -1,7 +1,9 @@
 snd-soc-core-objs := soc-core.o soc-dapm.o soc-jack.o soc-cache.o soc-utils.o
 snd-soc-core-objs += soc-pcm.o soc-io.o soc-devres.o soc-ops.o
 snd-soc-core-objs += soc-topology.o
-snd-soc-core-$(CONFIG_SND_SOC_COMPRESS) += soc-compress.o
+snd-soc-compress-objs := soc-compress.o
+
+obj-$(CONFIG_SND_SOC_COMPRESS) += snd-soc-compress.o
 
 ifneq ($(CONFIG_SND_SOC_GENERIC_DMAENGINE_PCM),)
 snd-soc-core-objs += soc-generic-dmaengine-pcm.o
sound/soc/soc-compress.c +8 -0
index 025c38f..a93ac8a 100644
@@ -19,6 +19,7 @@
 #include <linux/delay.h>
 #include <linux/slab.h>
 #include <linux/workqueue.h>
+#include <linux/module.h>
 #include <sound/core.h>
 #include <sound/compress_params.h>
 #include <sound/compress_driver.h>
@@ -703,3 +704,10 @@ compr_err:
 	kfree(compr);
 	return ret;
 }
+EXPORT_SYMBOL_GPL(soc_new_compress);
+
+/* Module information */
+MODULE_AUTHOR("Ramesh Babu K V <[email protected]>");
+MODULE_AUTHOR("Vinod Koul <[email protected]>");
+MODULE_DESCRIPTION("ALSA SoC Compress");
+MODULE_LICENSE("GPL v2");
Signature · 1 line
1.9.1
maintainer
#2 lore raw

Re: [PATCH v2 2/2] ASoC: soc-compress: split soc-compress to a module

Quoted 2 lines · Jie Yang

Split soc-compress into a separate module so we only compile/load it when it's needed.

Given that this patch is so much simpler than the previous one it would seem better to make it the first patch - I'd be OK to apply it but it depends on the prior patch.

#3 lore raw
In reply to Jie Yang

At Mon, 15 Jun 2015 11:20:44 +0800,

Quoted 19 lines · Jie Yang

We don't always need soc-compress in soc, here add a config item SND_SOC_COMPRESS, please add 'select SND_SOC_COMPRESS' to driver Kconfig when it is needed.

---
 include/sound/soc.h     | 7 +++++++
 sound/soc/Kconfig       | 5 ++++-
 sound/soc/Makefile      | 3 ++-
 sound/soc/intel/Kconfig | 1 +
 4 files changed, 14 insertions(+), 2 deletions(-)
include/sound/soc.h +0 -0
index 93df8bf..cc1cd4f 100644
@@ -440,7 +440,14 @@ int snd_soc_platform_read(struct snd_soc_platform *platform,
 int snd_soc_platform_write(struct snd_soc_platform *platform,

include/sound/soc.h int snd_soc_platform_read(struct snd_soc_platform *platform, · line 450

Quoted 10 lines
include/sound/soc.h +7 -0
 					unsigned int reg, unsigned int val);
 int soc_new_pcm(struct snd_soc_pcm_runtime *rtd, int num);
+#if IS_ENABLED(CONFIG_SND_SOC_COMPRESS)
 int soc_new_compress(struct snd_soc_pcm_runtime *rtd, int num);
+#else
+static inline int soc_new_compress(struct snd_soc_pcm_runtime *rtd, int num)
+{
+       return -EPERM;
+}
+#endif

A dummy function in such a case has both merit and demerit. The demerit is that you won't get errors if the driver really wants the compress support but just forgot to select the dependency.

Takashi

Quoted 53 lines
include/sound/soc.h +0 -0
 struct snd_pcm_substream *snd_soc_get_dai_substream(struct snd_soc_card *card,
 		const char *dai_link, int stream);
sound/soc/Kconfig +4 -1
index e2828e1..a124759 100644
@@ -9,7 +9,6 @@ menuconfig SND_SOC
 	select SND_JACK if INPUT=y || INPUT=SND
 	select REGMAP_I2C if I2C
 	select REGMAP_SPI if SPI_MASTER
-	select SND_COMPRESS_OFFLOAD
 	---help---
 
 	  If you want ASoC support, you should say Y here and also to the
@@ -30,6 +29,10 @@ config SND_SOC_GENERIC_DMAENGINE_PCM
 	bool
 	select SND_DMAENGINE_PCM
 
+config SND_SOC_COMPRESS
+	tristate
+	select SND_COMPRESS_OFFLOAD
+
 # All the supported SoCs
 source "sound/soc/adi/Kconfig"
 source "sound/soc/atmel/Kconfig"
sound/soc/Makefile +2 -1
index a0e1ee6..184c1e6 100644
@@ -1,6 +1,7 @@
 snd-soc-core-objs := soc-core.o soc-dapm.o soc-jack.o soc-cache.o soc-utils.o
-snd-soc-core-objs += soc-pcm.o soc-compress.o soc-io.o soc-devres.o soc-ops.o
+snd-soc-core-objs += soc-pcm.o soc-io.o soc-devres.o soc-ops.o
 snd-soc-core-objs += soc-topology.o
+snd-soc-core-$(CONFIG_SND_SOC_COMPRESS) += soc-compress.o
 
 ifneq ($(CONFIG_SND_SOC_GENERIC_DMAENGINE_PCM),)
 snd-soc-core-objs += soc-generic-dmaengine-pcm.o
sound/soc/intel/Kconfig +1 -0
index 791953f..e559174 100644
@@ -12,6 +12,7 @@ config SND_MFLD_MACHINE
 
 config SND_SST_MFLD_PLATFORM
 	tristate
+	select SND_SOC_COMPRESS
 
 config SND_SST_IPC
 	tristate
Signature · 1 line
1.9.1
author
#4 lore raw
Quoted 33 lines

-----Original Message-----
From: Takashi Iwai [mailto:[email protected]]
Sent: Monday, June 15, 2015 7:34 PM
To: Jie, Yang
Cc: [email protected]; [email protected]; Girdwood, Liam R; [email protected]; [email protected]; Zhang, Vivian
Subject: Re: [PATCH v2 1/2] ASoC: soc-compress: add a config item for soc- compress

At Mon, 15 Jun 2015 11:20:44 +0800,

include/sound/soc.h int snd_soc_platform_read(struct

Quoted 18 lines · Jie Yang

We don't always need soc-compress in soc, here add a config item SND_SOC_COMPRESS, please add 'select SND_SOC_COMPRESS' to driver Kconfig when it is needed.

---
 include/sound/soc.h     | 7 +++++++
 sound/soc/Kconfig       | 5 ++++-
 sound/soc/Makefile      | 3 ++-
 sound/soc/intel/Kconfig | 1 +
 4 files changed, 14 insertions(+), 2 deletions(-)
include/sound/soc.h index renamed from include/sound/soc.h +0 -0

93df8bf..cc1cd4f 100644

include/sound/soc.h +0 -0
@@ -440,7 +440,14 @@ int snd_soc_platform_read(struct

snd_soc_platform

include/sound/soc.h int snd_soc_platform_read(struct · line 441

Quoted 2 lines
include/sound/soc.h +0 -0
 *platform,  int snd_soc_platform_write(struct snd_soc_platform *platform,
 					unsigned int reg, unsigned int val);

int soc_new_pcm(struct

Quoted 13 lines

include/sound/soc.h int snd_soc_platform_read(struct snd_soc_platform *platform, · line 450

Quoted 9 lines
include/sound/soc.h +7 -0
 snd_soc_pcm_runtime *rtd, int num);
+#if IS_ENABLED(CONFIG_SND_SOC_COMPRESS)
 int soc_new_compress(struct snd_soc_pcm_runtime *rtd, int num);
+#else
+static inline int soc_new_compress(struct snd_soc_pcm_runtime *rtd,
+int num) {
+       return -EPERM;
+}
+#endif

A dummy function in such a case has both merit and demerit. The demerit is that you won't get errors if the driver really wants the compress support but just forgot to select the dependency.

Here I used -EPERM to return and tell caller that compress operation is not permitted, does it make sense, Takashi?

Or can you suggest a better solution, according to this specific case?

Thanks,
~Keyon

Quoted 59 lines

Takashi

include/sound/soc.h int snd_soc_platform_read(struct · line 451

Quoted 1 line
include/sound/soc.h +0 -0
 struct snd_pcm_substream *snd_soc_get_dai_substream(struct

snd_soc_card *card,

Quoted 53 lines
include/sound/soc.h +0 -0
 		const char *dai_link, int stream);
sound/soc/Kconfig index renamed from sound/soc/Kconfig +0 -0

e2828e1..a124759 100644

sound/soc/Kconfig +4 -2
@@ -9,7 +9,6 @@ menuconfig SND_SOC
 	select SND_JACK if INPUT=y || INPUT=SND
 	select REGMAP_I2C if I2C
 	select REGMAP_SPI if SPI_MASTER
-	select SND_COMPRESS_OFFLOAD
 	---help---
 
 	  If you want ASoC support, you should say Y here and also to the @@
-30,6 +29,10 @@ config SND_SOC_GENERIC_DMAENGINE_PCM
 	bool
 	select SND_DMAENGINE_PCM
 
+config SND_SOC_COMPRESS
+	tristate
+	select SND_COMPRESS_OFFLOAD
+
 # All the supported SoCs
 source "sound/soc/adi/Kconfig"
 source "sound/soc/atmel/Kconfig"
sound/soc/Makefile index renamed from sound/soc/Makefile +0 -0

a0e1ee6..184c1e6 100644

sound/soc/Makefile +0 -0
@@ -1,6 +1,7 @@
 snd-soc-core-objs := soc-core.o soc-dapm.o soc-jack.o soc-cache.o

soc-utils.o -snd-soc-core-objs += soc-pcm.o soc-compress.o soc-io.o soc-devres.o soc-ops.o

sound/soc/Makefile +2 -0
+snd-soc-core-objs += soc-pcm.o soc-io.o soc-devres.o soc-ops.o
 snd-soc-core-objs += soc-topology.o
+snd-soc-core-$(CONFIG_SND_SOC_COMPRESS) += soc-compress.o
 
 ifneq ($(CONFIG_SND_SOC_GENERIC_DMAENGINE_PCM),)
 snd-soc-core-objs += soc-generic-dmaengine-pcm.o diff --git

a/sound/soc/intel/Kconfig b/sound/soc/intel/Kconfig index 791953f..e559174 100644

sound/soc/intel/Kconfig +1 -1
@@ -12,6 +12,7 @@ config SND_MFLD_MACHINE
 
 config SND_SST_MFLD_PLATFORM
 	tristate
+	select SND_SOC_COMPRESS
 
 config SND_SST_IPC
 	tristate
--

1.9.1

#5 lore raw

At Mon, 15 Jun 2015 14:15:40 +0000,

Quoted 48 lines · Yang
Quoted 34 lines

-----Original Message-----
From: Takashi Iwai [mailto:[email protected]]
Sent: Monday, June 15, 2015 7:34 PM
To: Jie, Yang
Cc: [email protected]; [email protected]; Girdwood, Liam R; [email protected]; [email protected]; Zhang, Vivian
Subject: Re: [PATCH v2 1/2] ASoC: soc-compress: add a config item for soc- compress

At Mon, 15 Jun 2015 11:20:44 +0800,

include/sound/soc.h int snd_soc_platform_read(struct

Quoted 18 lines · Jie Yang

We don't always need soc-compress in soc, here add a config item SND_SOC_COMPRESS, please add 'select SND_SOC_COMPRESS' to driver Kconfig when it is needed.

---
 include/sound/soc.h     | 7 +++++++
 sound/soc/Kconfig       | 5 ++++-
 sound/soc/Makefile      | 3 ++-
 sound/soc/intel/Kconfig | 1 +
 4 files changed, 14 insertions(+), 2 deletions(-)
include/sound/soc.h index renamed from include/sound/soc.h +0 -0

93df8bf..cc1cd4f 100644

include/sound/soc.h +0 -0
@@ -440,7 +440,14 @@ int snd_soc_platform_read(struct

snd_soc_platform

include/sound/soc.h int snd_soc_platform_read(struct · line 441

Quoted 2 lines
include/sound/soc.h +0 -0
 *platform,  int snd_soc_platform_write(struct snd_soc_platform *platform,
 					unsigned int reg, unsigned int val);

int soc_new_pcm(struct

Quoted 13 lines

include/sound/soc.h int snd_soc_platform_read(struct snd_soc_platform *platform, · line 450

Quoted 9 lines
include/sound/soc.h +7 -0
 snd_soc_pcm_runtime *rtd, int num);
+#if IS_ENABLED(CONFIG_SND_SOC_COMPRESS)
 int soc_new_compress(struct snd_soc_pcm_runtime *rtd, int num);
+#else
+static inline int soc_new_compress(struct snd_soc_pcm_runtime *rtd,
+int num) {
+       return -EPERM;
+}
+#endif

A dummy function in such a case has both merit and demerit. The demerit is that you won't get errors if the driver really wants the compress support but just forgot to select the dependency.

Quoted 2 lines

Here I used -EPERM to return and tell caller that compress operation is not permitted, does it make sense, Takashi?

The question is whether a runtime error is the best option. A runtime error won't be caught by build tests but only when actually running by a user.

Quoted 1 line

Or can you suggest a better solution, according to this specific case?

Well, there are several ways, obviously (e.g. give the preprocessor error if CONFIG_SND_SOC_COMPRESS is enabled). But I'm not sure what is the best to fit, too. It's rather a matter of taste.

Takashi

Quoted 62 lines

Thanks,
~Keyon

Quoted 59 lines

Takashi

include/sound/soc.h int snd_soc_platform_read(struct · line 451

Quoted 1 line
include/sound/soc.h +0 -0
 struct snd_pcm_substream *snd_soc_get_dai_substream(struct

snd_soc_card *card,

Quoted 53 lines
include/sound/soc.h +0 -0
 		const char *dai_link, int stream);
sound/soc/Kconfig index renamed from sound/soc/Kconfig +0 -0

e2828e1..a124759 100644

sound/soc/Kconfig +4 -2
@@ -9,7 +9,6 @@ menuconfig SND_SOC
 	select SND_JACK if INPUT=y || INPUT=SND
 	select REGMAP_I2C if I2C
 	select REGMAP_SPI if SPI_MASTER
-	select SND_COMPRESS_OFFLOAD
 	---help---
 
 	  If you want ASoC support, you should say Y here and also to the @@
-30,6 +29,10 @@ config SND_SOC_GENERIC_DMAENGINE_PCM
 	bool
 	select SND_DMAENGINE_PCM
 
+config SND_SOC_COMPRESS
+	tristate
+	select SND_COMPRESS_OFFLOAD
+
 # All the supported SoCs
 source "sound/soc/adi/Kconfig"
 source "sound/soc/atmel/Kconfig"
sound/soc/Makefile index renamed from sound/soc/Makefile +0 -0

a0e1ee6..184c1e6 100644

sound/soc/Makefile +0 -0
@@ -1,6 +1,7 @@
 snd-soc-core-objs := soc-core.o soc-dapm.o soc-jack.o soc-cache.o

soc-utils.o -snd-soc-core-objs += soc-pcm.o soc-compress.o soc-io.o soc-devres.o soc-ops.o

sound/soc/Makefile +2 -0
+snd-soc-core-objs += soc-pcm.o soc-io.o soc-devres.o soc-ops.o
 snd-soc-core-objs += soc-topology.o
+snd-soc-core-$(CONFIG_SND_SOC_COMPRESS) += soc-compress.o
 
 ifneq ($(CONFIG_SND_SOC_GENERIC_DMAENGINE_PCM),)
 snd-soc-core-objs += soc-generic-dmaengine-pcm.o diff --git

a/sound/soc/intel/Kconfig b/sound/soc/intel/Kconfig index 791953f..e559174 100644

sound/soc/intel/Kconfig +1 -1
@@ -12,6 +12,7 @@ config SND_MFLD_MACHINE
 
 config SND_SST_MFLD_PLATFORM
 	tristate
+	select SND_SOC_COMPRESS
 
 config SND_SST_IPC
 	tristate
--

1.9.1

author
#6 lore raw
Quoted 62 lines

-----Original Message-----
From: Takashi Iwai [mailto:[email protected]]
Sent: Monday, June 15, 2015 10:27 PM
To: Jie, Yang
Cc: [email protected]; [email protected]; Girdwood, Liam R; [email protected]; [email protected]; Zhang, Vivian
Subject: Re: [PATCH v2 1/2] ASoC: soc-compress: add a config item for soc- compress

At Mon, 15 Jun 2015 14:15:40 +0000,

Quoted 11 lines · Yang
Quoted 11 lines

-----Original Message-----
From: Takashi Iwai [mailto:[email protected]]
Sent: Monday, June 15, 2015 7:34 PM
To: Jie, Yang
Cc: [email protected]; [email protected]; Girdwood, Liam R; [email protected]; [email protected]; Zhang, Vivian
Subject: Re: [PATCH v2 1/2] ASoC: soc-compress: add a config item for soc- compress

Quoted 5 lines
Quoted 5 lines

At Mon, 15 Jun 2015 11:20:44 +0800,

Quoted 3 lines · Jie Yang

We don't always need soc-compress in soc, here add a config item SND_SOC_COMPRESS, please add 'select SND_SOC_COMPRESS' to

driver

Quoted 10 lines
Quoted 10 lines
Quoted 10 lines

Kconfig when it is needed.

---
 include/sound/soc.h     | 7 +++++++
 sound/soc/Kconfig       | 5 ++++-
 sound/soc/Makefile      | 3 ++-
 sound/soc/intel/Kconfig | 1 +
 4 files changed, 14 insertions(+), 2 deletions(-)
Quoted 7 lines
Quoted 7 lines

include/sound/soc.h int snd_soc_platform_read(struct

Quoted 5 lines
include/sound/soc.h index renamed from include/sound/soc.h +0 -0

93df8bf..cc1cd4f 100644

include/sound/soc.h +0 -0
@@ -440,7 +440,14 @@ int snd_soc_platform_read(struct

snd_soc_platform

Quoted 1 line
include/sound/soc.h +0 -0
 *platform,  int snd_soc_platform_write(struct snd_soc_platform

*platform,

Quoted 16 lines
Quoted 15 lines

include/sound/soc.h int snd_soc_platform_read(struct snd_soc_platform *platform, · line 441

Quoted 1 line
include/sound/soc.h +0 -0
 					unsigned int reg, unsigned int val);

int soc_new_pcm(struct

include/sound/soc.h int snd_soc_platform_read(struct snd_soc_platform *platform, · line 450

Quoted 9 lines
include/sound/soc.h +7 -0
 snd_soc_pcm_runtime *rtd, int num);
+#if IS_ENABLED(CONFIG_SND_SOC_COMPRESS)
 int soc_new_compress(struct snd_soc_pcm_runtime *rtd, int num);
+#else
+static inline int soc_new_compress(struct snd_soc_pcm_runtime
+*rtd, int num) {
+       return -EPERM;
+}
+#endif

A dummy function in such a case has both merit and demerit. The demerit is that you won't get errors if the driver really wants the compress support but just forgot to select the dependency.

Quoted 5 lines
Quoted 2 lines

Here I used -EPERM to return and tell caller that compress operation is not permitted, does it make sense, Takashi?

The question is whether a runtime error is the best option. A runtime error won't be caught by build tests but only when actually running by a user.

Unfortunately, here cpu_dai->driver->compress_dai is a runtime value, which means we need compress API when it is true. Seems it is not easy to decide it at compile stage?

Quoted 5 lines
Quoted 2 lines

Or can you suggest a better solution, according to this specific case?

Well, there are several ways, obviously (e.g. give the preprocessor error if CONFIG_SND_SOC_COMPRESS is enabled). But I'm not sure what is the best

Do you mean giving error, when CONFIG_SND_SOC_COMPRESS is disabled but the driver wants the compress support at the same time?

As I stated above, it's not easy to detect whether the driver wants the compress support, at the preprocessor/compile stage. :(

~Keyon

Quoted 68 lines

to fit, too. It's rather a matter of taste.

Takashi

Quoted 63 lines

Thanks,
~Keyon

Quoted 60 lines

Takashi

include/sound/soc.h int snd_soc_platform_read(struct · line 451

Quoted 1 line
include/sound/soc.h +0 -0
 struct snd_pcm_substream *snd_soc_get_dai_substream(struct

snd_soc_card *card,

Quoted 54 lines
include/sound/soc.h +0 -0
 		const char *dai_link, int stream); diff --git

a/sound/soc/Kconfig b/sound/soc/Kconfig index
e2828e1..a124759 100644

sound/soc/Kconfig +0 -1
@@ -9,7 +9,6 @@ menuconfig SND_SOC
 	select SND_JACK if INPUT=y || INPUT=SND
 	select REGMAP_I2C if I2C
 	select REGMAP_SPI if SPI_MASTER
-	select SND_COMPRESS_OFFLOAD
 	---help---
 
 	  If you want ASoC support, you should say Y here and also to

the @@
-30,6 +29,10 @@ config SND_SOC_GENERIC_DMAENGINE_PCM

sound/soc/Kconfig +4 -0
 	bool
 	select SND_DMAENGINE_PCM
 
+config SND_SOC_COMPRESS
+	tristate
+	select SND_COMPRESS_OFFLOAD
+
 # All the supported SoCs
 source "sound/soc/adi/Kconfig"
 source "sound/soc/atmel/Kconfig"
sound/soc/Makefile index renamed from sound/soc/Makefile +0 -0

a0e1ee6..184c1e6 100644

sound/soc/Makefile +0 -0
@@ -1,6 +1,7 @@
 snd-soc-core-objs := soc-core.o soc-dapm.o soc-jack.o soc-cache.o

soc-utils.o -snd-soc-core-objs += soc-pcm.o soc-compress.o soc-io.o soc-devres.o soc-ops.o

sound/soc/Makefile +2 -0
+snd-soc-core-objs += soc-pcm.o soc-io.o soc-devres.o soc-ops.o
 snd-soc-core-objs += soc-topology.o
+snd-soc-core-$(CONFIG_SND_SOC_COMPRESS) += soc-compress.o
 
 ifneq ($(CONFIG_SND_SOC_GENERIC_DMAENGINE_PCM),)
 snd-soc-core-objs += soc-generic-dmaengine-pcm.o diff --git

a/sound/soc/intel/Kconfig b/sound/soc/intel/Kconfig index 791953f..e559174 100644

sound/soc/intel/Kconfig +1 -1
@@ -12,6 +12,7 @@ config SND_MFLD_MACHINE
 
 config SND_SST_MFLD_PLATFORM
 	tristate
+	select SND_SOC_COMPRESS
 
 config SND_SST_IPC
 	tristate
--

1.9.1

maintainer
#7 lore raw
Quoted 9 lines · Yang
Quoted 5 lines
Quoted 2 lines

Here I used -EPERM to return and tell caller that compress operation is not permitted, does it make sense, Takashi?

The question is whether a runtime error is the best option. A runtime error won't be caught by build tests but only when actually running by a user.

Unfortunately, here cpu_dai->driver->compress_dai is a runtime value, which means we need compress API when it is true. Seems it is not easy to decide it at compile stage?

The machine driver (which is presumably the thing that should be doing the select here, it's not user visible) really ought to know if the DAI links it is creating are for compressed audio. This is why I'm still surprised there's no driver updates as part of this patch. If the Intel drivers referencing compressed don't actually implement it then I'd expect to see patches cleaning up the references to compressed audio.

author
#8 lore raw
Quoted 22 lines

-----Original Message-----
From: Mark Brown [mailto:[email protected]]
Sent: Monday, June 15, 2015 11:06 PM
To: Jie, Yang
Cc: Takashi Iwai; [email protected]; Girdwood, Liam R; [email protected]; [email protected]; Zhang, Vivian
Subject: Re: [PATCH v2 1/2] ASoC: soc-compress: add a config item for soc- compress

Quoted 5 lines · Yang
Quoted 5 lines
Quoted 2 lines

Here I used -EPERM to return and tell caller that compress operation is not permitted, does it make sense, Takashi?

The question is whether a runtime error is the best option. A runtime error won't be caught by build tests but only when actually

running by a user.

Quoted 3 lines

Unfortunately, here cpu_dai->driver->compress_dai is a runtime value, which means we need compress API when it is true. Seems it is not easy to decide it at compile stage?

Quoted 6 lines

The machine driver (which is presumably the thing that should be doing the select here, it's not user visible) really ought to know if the DAI links it is creating are for compressed audio. This is why I'm still surprised there's no driver updates as part of this patch. If the Intel drivers referencing compressed don't actually implement it then I'd expect to see patches cleaning up the references to compressed audio.

Mark, as you may see in this patch(1/2), I have added this selection to sound/ Soc/intel/Kconfig, we can find .compress_dai =1 in sst-mfld-platform-pcm.c, the compressed is referenced in sst-mfld-platform-pcm.c:sst_platform_dai, here copy the related part from the patch:

b/sound/soc/intel/Kconfig index 791953f..e559174 100644

sound/soc/intel/Kconfig +1 -0
@@ -12,6 +12,7 @@ config SND_MFLD_MACHINE
 
 config SND_SST_MFLD_PLATFORM
 	tristate
+	select SND_SOC_COMPRESS
 
 config SND_SST_IPC
 	tristate

~Keyon

maintainer
#9 lore raw
Quoted 4 lines · Yang

Mark, as you may see in this patch(1/2), I have added this selection to sound/ Soc/intel/Kconfig, we can find .compress_dai =1 in sst-mfld-platform-pcm.c, the compressed is referenced in sst-mfld-platform-pcm.c:sst_platform_dai, here copy the related part from the patch:

OK, that should have been in the commit message.

author
#10 lore raw
Quoted 11 lines

-----Original Message-----
From: Mark Brown [mailto:[email protected]]
Sent: Tuesday, June 16, 2015 6:54 PM
To: Jie, Yang
Cc: Takashi Iwai; [email protected]; Girdwood, Liam R; [email protected]; [email protected]; Zhang, Vivian
Subject: Re: [PATCH v2 1/2] ASoC: soc-compress: add a config item for soc- compress

On Tue, Jun 16, 2015 at 12:49:37AM +0000, Jie, Yang wrote:

Quoted 7 lines
Quoted 5 lines

Mark, as you may see in this patch(1/2), I have added this selection to sound/ Soc/intel/Kconfig, we can find .compress_dai =1 in sst-mfld-platform-pcm.c, the compressed is referenced in sst-mfld-platform-pcm.c:sst_platform_dai,
here copy the related part from the patch:

OK, that should have been in the commit message.

OK, so let me add it to the commit message and resend?

~Keyon

maintainer
#11 lore raw
Quoted 3 lines · Yang
Quoted 1 line

OK, that should have been in the commit message.

OK, so let me add it to the commit message and resend?

Well, there's still the stubs to consider. Should we really be returning an error or should we silently ignore the error and hide the DAI if the user deconfigured compressed audio? Or rearrange things so we don't need stubs? Given that the machine driver has to select compressed support it's not something that should ever really hit the stubs, it's not truly a runtime thing...

#12 lore raw

At Tue, 16 Jun 2015 17:06:07 +0100,

Quoted 5 lines · Mark Brown
Quoted 1 line · Yang
Quoted 1 line

OK, that should have been in the commit message.

Quoted 8 lines
Quoted 1 line

OK, so let me add it to the commit message and resend?

Well, there's still the stubs to consider. Should we really be returning an error or should we silently ignore the error and hide the DAI if the user deconfigured compressed audio? Or rearrange things so we don't need stubs? Given that the machine driver has to select compressed support it's not something that should ever really hit the stubs, it's not truly a runtime thing...

Yes, I guess that leaving without dummy function will give unresolved symbol errors at module link time, so the user can catch before actually running it. Of course, this should be checked actually.

Takashi

#13 lore raw

At Tue, 16 Jun 2015 18:14:40 +0200,

Quoted 17 lines · Takashi Iwai

At Tue, 16 Jun 2015 17:06:07 +0100,

Quoted 5 lines · Mark Brown
Quoted 1 line · Yang
Quoted 1 line

OK, that should have been in the commit message.

Quoted 8 lines
Quoted 1 line

OK, so let me add it to the commit message and resend?

Well, there's still the stubs to consider. Should we really be returning an error or should we silently ignore the error and hide the DAI if the user deconfigured compressed audio? Or rearrange things so we don't need stubs? Given that the machine driver has to select compressed support it's not something that should ever really hit the stubs, it's not truly a runtime thing...

Quoted 3 lines

Yes, I guess that leaving without dummy function will give unresolved symbol errors at module link time, so the user can catch before actually running it. Of course, this should be checked actually.

Oh, scratch this. It won't work well in the current code. Possibly with weak linking, but...

Takashi

#14 lore raw
Quoted 19 lines · Takashi Iwai

At Tue, 16 Jun 2015 18:14:40 +0200,

Quoted 17 lines · Takashi Iwai

At Tue, 16 Jun 2015 17:06:07 +0100,

Quoted 13 lines · Mark Brown
Quoted 3 lines · Yang
Quoted 1 line

OK, that should have been in the commit message.

OK, so let me add it to the commit message and resend?

Well, there's still the stubs to consider. Should we really be returning an error or should we silently ignore the error and hide the DAI if the user deconfigured compressed audio? Or rearrange things so we don't need stubs? Given that the machine driver has to select compressed support it's not something that should ever really hit the stubs, it's not truly a runtime thing...

Quoted 6 lines
Quoted 3 lines

Yes, I guess that leaving without dummy function will give unresolved symbol errors at module link time, so the user can catch before actually running it. Of course, this should be checked actually.

Oh, scratch this. It won't work well in the current code. Possibly with weak linking, but...

Right as the soc_new_compress() is called from soc-core for compressed dai links. Since ASoC drivers don't use any of compress APIs directly, it would be difficult to add a compile time failure so we are stuck with silent runtime failures :(

Signature · 1 line
~Vinod
#15 lore raw

At Tue, 16 Jun 2015 21:55:07 +0530,

Quoted 25 lines · Vinod Koul
Quoted 23 lines · Takashi Iwai

At Tue, 16 Jun 2015 18:14:40 +0200,

Quoted 20 lines · Takashi Iwai

At Tue, 16 Jun 2015 17:06:07 +0100,

Quoted 5 lines · Mark Brown
Quoted 1 line · Yang
Quoted 1 line

OK, that should have been in the commit message.

Quoted 8 lines
Quoted 1 line

OK, so let me add it to the commit message and resend?

Well, there's still the stubs to consider. Should we really be returning an error or should we silently ignore the error and hide the DAI if the user deconfigured compressed audio? Or rearrange things so we don't need stubs? Given that the machine driver has to select compressed support it's not something that should ever really hit the stubs, it's not truly a runtime thing...

Yes, I guess that leaving without dummy function will give unresolved symbol errors at module link time, so the user can catch before actually running it. Of course, this should be checked actually.

Quoted 6 lines
Quoted 2 lines

Oh, scratch this. It won't work well in the current code. Possibly with weak linking, but...

Right as the soc_new_compress() is called from soc-core for compressed dai links. Since ASoC drivers don't use any of compress APIs directly, it would be difficult to add a compile time failure so we are stuck with silent runtime failures :(

One feasible option would be to make the driver's dai creation a function pointer to the generic constructor. That is, soc_probe_link_dais() will run like:

	if (cpu_dai->driver->create)
		ret = cpu_dai->driver->create(rtd, num);

and each driver passes soc_new_compress (or whatever) there. This is flexible and can be extended if any new stuff has to be handled. Even PCM creation or dai widget links can be passed in that form, too.

Takashi

#16 lore raw
Quoted 38 lines · Takashi Iwai

At Tue, 16 Jun 2015 21:55:07 +0530,

Quoted 25 lines · Vinod Koul
Quoted 23 lines · Takashi Iwai

At Tue, 16 Jun 2015 18:14:40 +0200,

Quoted 20 lines · Takashi Iwai

At Tue, 16 Jun 2015 17:06:07 +0100,

Quoted 5 lines · Mark Brown
Quoted 1 line · Yang
Quoted 1 line

OK, that should have been in the commit message.

Quoted 8 lines
Quoted 1 line

OK, so let me add it to the commit message and resend?

Well, there's still the stubs to consider. Should we really be returning an error or should we silently ignore the error and hide the DAI if the user deconfigured compressed audio? Or rearrange things so we don't need stubs? Given that the machine driver has to select compressed support it's not something that should ever really hit the stubs, it's not truly a runtime thing...

Yes, I guess that leaving without dummy function will give unresolved symbol errors at module link time, so the user can catch before actually running it. Of course, this should be checked actually.

Quoted 6 lines
Quoted 2 lines

Oh, scratch this. It won't work well in the current code. Possibly with weak linking, but...

Right as the soc_new_compress() is called from soc-core for compressed dai links. Since ASoC drivers don't use any of compress APIs directly, it would be difficult to add a compile time failure so we are stuck with silent runtime failures :(

One feasible option would be to make the driver's dai creation a function pointer to the generic constructor. That is, soc_probe_link_dais() will run like:

Quoted 7 lines
	if (cpu_dai->driver->create)
		ret = cpu_dai->driver->create(rtd, num);

and each driver passes soc_new_compress (or whatever) there. This is flexible and can be extended if any new stuff has to be handled. Even PCM creation or dai widget links can be passed in that form, too.

and the default would be snd_pcm_new so that *most* drivers dont need to add this. Only Intel atom driver should add soc_new_compress() as create ops, hence direct reference so no gaurds required

Looks good to me then, Mark you okay with this approach?

Signature · 1 line
~Vinod
maintainer
#17 lore raw
Quoted 1 line · Vinod Koul

On Tue, Jun 16, 2015 at 06:36:22PM +0200, Takashi Iwai wrote:

Please remember to delete unneeded context from replies.

Quoted 9 lines
Quoted 8 lines
	if (cpu_dai->driver->create)
		ret = cpu_dai->driver->create(rtd, num);

and each driver passes soc_new_compress (or whatever) there. This is flexible and can be extended if any new stuff has to be handled. Even PCM creation or dai widget links can be passed in that form, too.

Quoted 5 lines

and the default would be snd_pcm_new so that *most* drivers dont need to add this. Only Intel atom driver should add soc_new_compress() as create ops, hence direct reference so no gaurds required

Looks good to me then, Mark you okay with this approach?

I'm not sure this fully helps. This will put the reference in the CPU driver which if there's sharing with non-compressed DAIs means that any machine driver using the platform needs to select the compressed audio code. It also means that we're still in the situation whre if a machine *can* support compressed audio it *must* support compressed audio, it's not clear to me that people doing this sort of memory optimisation are always going to be doing it on hardware that doesn't have the capacity for compressed audio (or won't in the future). But then I don't know exactly how much memory is being saved here...

#18 lore raw
Quoted 16 lines · Mark Brown
Quoted 9 lines · Vinod Koul
Quoted 8 lines
	if (cpu_dai->driver->create)
		ret = cpu_dai->driver->create(rtd, num);

and each driver passes soc_new_compress (or whatever) there. This is flexible and can be extended if any new stuff has to be handled. Even PCM creation or dai widget links can be passed in that form, too.

Quoted 5 lines

and the default would be snd_pcm_new so that *most* drivers dont need to add this. Only Intel atom driver should add soc_new_compress() as create ops, hence direct reference so no gaurds required

Looks good to me then, Mark you okay with this approach?

Quoted 9 lines

I'm not sure this fully helps. This will put the reference in the CPU driver which if there's sharing with non-compressed DAIs means that any machine driver using the platform needs to select the compressed audio code. It also means that we're still in the situation whre if a machine *can* support compressed audio it *must* support compressed audio, it's not clear to me that people doing this sort of memory optimisation are always going to be doing it on hardware that doesn't have the capacity for compressed audio (or won't in the future). But then I don't know exactly how much memory is being saved here...

Yes this is a good point but then driver can define the compressed dai only when SND_SOC_COMPRESS is enabled. So if machine has enabled this symbol then only dai get added and machine can create dai-link

Also for the size question, in sound-next we have 248KB soc-compress.o

Thanks

Signature · 1 line
~Vinod
maintainer
#19 lore raw
Quoted 12 lines · Vinod Koul
Quoted 9 lines · Mark Brown

I'm not sure this fully helps. This will put the reference in the CPU driver which if there's sharing with non-compressed DAIs means that any machine driver using the platform needs to select the compressed audio code. It also means that we're still in the situation whre if a machine *can* support compressed audio it *must* support compressed audio, it's not clear to me that people doing this sort of memory optimisation are always going to be doing it on hardware that doesn't have the capacity for compressed audio (or won't in the future). But then I don't know exactly how much memory is being saved here...

Quoted 3 lines

Yes this is a good point but then driver can define the compressed dai only when SND_SOC_COMPRESS is enabled. So if machine has enabled this symbol then only dai get added and machine can create dai-link

Well, if it's a user configurable thing then the stubs start to make sense - the driver defines the DAI and then the core doesn't bother to instantiate it.

Quoted 1 line

Also for the size question, in sound-next we have 248KB soc-compress.o

How big is the actual code - that sounds like it's got the debug symbols? I suppose I should go look myself...

author
#20 lore raw
In reply to Mark Brown
Quoted 30 lines

-----Original Message-----
From: [email protected] [mailto:alsa-devel- [email protected]] On Behalf Of Mark Brown
Sent: Wednesday, June 17, 2015 8:24 PM
To: Koul, Vinod
Cc: Takashi Iwai; [email protected]; Zhang, Vivian; Babu, Ramesh; Girdwood, Liam R
Subject: Re: [alsa-devel] [PATCH v2 1/2] ASoC: soc-compress: add a config item for soc-compress

Quoted 1 line · Vinod Koul

On Tue, Jun 16, 2015 at 06:36:22PM +0200, Takashi Iwai wrote:

Please remember to delete unneeded context from replies.

Quoted 9 lines
Quoted 8 lines
	if (cpu_dai->driver->create)
		ret = cpu_dai->driver->create(rtd, num);

and each driver passes soc_new_compress (or whatever) there. This is flexible and can be extended if any new stuff has to be handled. Even PCM creation or dai widget links can be passed in that form, too.

Quoted 5 lines

and the default would be snd_pcm_new so that *most* drivers dont need to add this. Only Intel atom driver should add soc_new_compress() as create ops, hence direct reference so no gaurds required

Looks good to me then, Mark you okay with this approach?

Quoted 9 lines

I'm not sure this fully helps. This will put the reference in the CPU driver which if there's sharing with non-compressed DAIs means that any machine driver using the platform needs to select the compressed audio code. It also means that we're still in the situation whre if a machine
*can* support compressed audio it *must* support compressed audio, it's not clear to me that people doing this sort of memory optimisation are always going to be doing it on hardware that doesn't have the capacity for compressed audio (or won't in the future). But then I don't know exactly how much memory is being saved here...

It saves ~12KB memory usage on i386 platform on my side.

~Keyon

author
#21 lore raw
In reply to Takashi Iwai
Quoted 47 lines

-----Original Message-----
From: Takashi Iwai [mailto:[email protected]]
Sent: Wednesday, June 17, 2015 12:36 AM
To: Koul, Vinod
Cc: Mark Brown; Jie, Yang; [email protected]; Girdwood, Liam R; [email protected]; Zhang, Vivian
Subject: Re: [PATCH v2 1/2] ASoC: soc-compress: add a config item for soc- compress

At Tue, 16 Jun 2015 21:55:07 +0530,

Quoted 20 lines · Vinod Koul
Quoted 18 lines · Takashi Iwai

At Tue, 16 Jun 2015 18:14:40 +0200,

Quoted 16 lines · Takashi Iwai
Quoted 13 lines · Mark Brown
Quoted 3 lines · Yang
Quoted 1 line

OK, that should have been in the commit message.

OK, so let me add it to the commit message and resend?

Well, there's still the stubs to consider. Should we really be returning an error or should we silently ignore the error and hide the DAI if the user deconfigured compressed audio? Or rearrange things so we don't need stubs? Given that the machine driver has to select compressed support it's not something that should ever really hit the stubs, it's not truly a runtime thing...

Quoted 3 lines
Quoted 3 lines
Quoted 3 lines

Yes, I guess that leaving without dummy function will give unresolved symbol errors at module link time, so the user can catch before actually running it. Of course, this should be checked

actually.

Quoted 7 lines
Quoted 3 lines

Oh, scratch this. It won't work well in the current code. Possibly with weak linking, but...

Right as the soc_new_compress() is called from soc-core for compressed dai links. Since ASoC drivers don't use any of compress APIs directly, it would be difficult to add a compile time failure so we are stuck with silent runtime failures :(

One feasible option would be to make the driver's dai creation a function pointer to the generic constructor. That is, soc_probe_link_dais() will run like:

Quoted 6 lines
	if (cpu_dai->driver->create)
		ret = cpu_dai->driver->create(rtd, num);

and each driver passes soc_new_compress (or whatever) there. This is flexible and can be extended if any new stuff has to be handled. Even PCM creation or dai widget links can be passed in that form, too.

Then we may need this in driver code:

static struct snd_soc_dai_driver sst_platform_dai[] = { ...,

{
	.name = "compress-cpu-dai",
	.create = soc_new_compress, //.compress_dai = 1,
//	.create = NULL, //.compress_dai = 0,
	.ops = &sst_compr_dai_ops,
	.playback = {
		.stream_name = "Compress Playback",
		.channels_min = SST_STEREO,
		.channels_max = SST_STEREO,
		.rates = SNDRV_PCM_RATE_48000,
		.formats = SNDRV_PCM_FMTBIT_S16_LE,
	},
},

}

With this implementation, for driver which needed compress, if we forget to add 'select SND_SOC_COMPRESS' in Kconfig of the driver, compiler will check and report error(no NULL inline function defined).

Sounds this feasible for me, any extra comment for it?

~Keyon

Quoted 3 lines

Takashi

author
#22 lore raw
In reply to Takashi Iwai
Quoted 47 lines

-----Original Message-----
From: Takashi Iwai [mailto:[email protected]]
Sent: Wednesday, June 17, 2015 12:36 AM
To: Koul, Vinod
Cc: Mark Brown; Jie, Yang; [email protected]; Girdwood, Liam R; [email protected]; Zhang, Vivian
Subject: Re: [PATCH v2 1/2] ASoC: soc-compress: add a config item for soc- compress

At Tue, 16 Jun 2015 21:55:07 +0530,

Quoted 20 lines · Vinod Koul
Quoted 18 lines · Takashi Iwai

At Tue, 16 Jun 2015 18:14:40 +0200,

Quoted 16 lines · Takashi Iwai
Quoted 13 lines · Mark Brown
Quoted 3 lines · Yang
Quoted 1 line

OK, that should have been in the commit message.

OK, so let me add it to the commit message and resend?

Well, there's still the stubs to consider. Should we really be returning an error or should we silently ignore the error and hide the DAI if the user deconfigured compressed audio? Or rearrange things so we don't need stubs? Given that the machine driver has to select compressed support it's not something that should ever really hit the stubs, it's not truly a runtime thing...

Quoted 3 lines
Quoted 3 lines
Quoted 3 lines

Yes, I guess that leaving without dummy function will give unresolved symbol errors at module link time, so the user can catch before actually running it. Of course, this should be checked

actually.

Quoted 7 lines
Quoted 3 lines

Oh, scratch this. It won't work well in the current code. Possibly with weak linking, but...

Right as the soc_new_compress() is called from soc-core for compressed dai links. Since ASoC drivers don't use any of compress APIs directly, it would be difficult to add a compile time failure so we are stuck with silent runtime failures :(

One feasible option would be to make the driver's dai creation a function pointer to the generic constructor. That is, soc_probe_link_dais() will run like:

Quoted 6 lines
	if (cpu_dai->driver->create)
		ret = cpu_dai->driver->create(rtd, num);

and each driver passes soc_new_compress (or whatever) there. This is flexible and can be extended if any new stuff has to be handled. Even PCM creation or dai widget links can be passed in that form, too.

How time flies! About 3 months passed away when I have chance to resume back to this thread.

Takashi, so as your suggestion, I will add a member to struct snd_soc_dai_driver:

struct snd_soc_dai_driver {
...
+ int (*create)(struct snd_soc_dai *dai);
- bool compress_dai;
...
}

And then change sst_platform_dai[] in sst-mfld-platform-pcm.c such as:

{
        .name = "compress-cpu-dai",
-        .compress_dai = 1,
+     .create = soc_new_compress,
        .ops = &sst_compr_dai_ops,
        .playback = {
                .stream_name = "Compress Playback",
                .channels_min = SST_STEREO,
                .channels_max = SST_STEREO,
                .rates = SNDRV_PCM_RATE_48000,
                .formats = SNDRV_PCM_FMTBIT_S16_LE,
        },
},

And then, we can remove the dummy soc_new_compress() inline definition, is that right?

~Keyon

Quoted 3 lines

Takashi

#23 lore raw
In reply to Takashi Iwai
Quoted 14 lines · Takashi Iwai

At Tue, 16 Jun 2015 18:14:40 +0200,

Quoted 12 lines · Takashi Iwai

At Tue, 16 Jun 2015 17:06:07 +0100,

Quoted 10 lines · Mark Brown
Quoted 2 lines · Yang
Quoted 1 line

OK, that should have been in the commit message.

OK, so let me add it to the commit message and resend?

Well, there's still the stubs to consider. Should we really be returning an error or should we silently ignore the error and hide the DAI if the user deconfigured compressed audio? Or rearrange things so we don't need stubs? Given that the machine driver has to select compressed support it's not something that should ever really hit the stubs, it's not truly a runtime thing...

Quoted 5 lines
Quoted 3 lines

Yes, I guess that leaving without dummy function will give unresolved symbol errors at module link time, so the user can catch before actually running it. Of course, this should be checked actually.

Oh, scratch this. It won't work well in the current code. Possibly with weak linking, but...

I don't know if I am missing something, but am I the only one who thinks these patches and the proposed changes are too complex for what seems to be little gain? Why is it that bad to always compile in soc-compress?

Qais

Quoted 6 lines

Takashi
_______________________________________________
Alsa-devel mailing list
[email protected]
http://mailman.alsa-project.org/mailman/listinfo/alsa-devel