ASoC: Remove module probe announcements from CODEC drivers
Subsystems (3)
PATCH 1/1 ASoC: Remove module probe announcements from CODEC drivers
--- sound/soc/codecs/ak4104.c | 1 - sound/soc/codecs/cs4270.c | 2 -- 2 files changed, 0 insertions(+), 3 deletions(-)
sound/soc/codecs/ak4104.c +0 -1
@@ -294,7 +294,6 @@ static struct spi_driver ak4104_spi_driver = {static int __init ak4104_init (void ) { -pr_info ("Asahi Kasei AK4104 ALSA SoC Codec Driver\n" );return spi_register_driver (&ak4104_spi_driver); }module_init (ak4104_init);
sound/soc/codecs/cs4270.c +0 -2
@@ -743,8 +743,6 @@ static struct i2c_driver cs4270_i2c_driver = {static int __init cs4270_init (void ) { -pr_info ("Cirrus Logic CS4270 ALSA SoC Codec Driver\n" ); -return i2c_add_driver (&cs4270_i2c_driver); }module_init (cs4270_init);
Signature · 1 line
1.7.2.3
sound/soc/codecs/cs4270.c static struct i2c_driver cs4270_i2c_driver = { · line 747
Quoted 6 lines · Mark Brown
static int
__initcs4270_init(void)
{sound/soc/codecs/cs4270.c +0
-2-pr_info ("Cirrus Logic CS4270 ALSA SoC Codec Driver\n" ); -return i2c_add_driver (&cs4270_i2c_driver); }
I'm not fond of this at all.
Can I get a justification for this change? On my boards, the CS4270 driver is compiled in-kernel, not as a module, and so when the kernel boots, the driver announces itself when it's loaded. I like this a lot, which is why I always put a pr_info() in the __init function of my modules.
Signature · 2 lines
Timur Tabi Linux kernel developer at Freescale
Quoted 4 lines · Timur Tabi
Can I get a justification for this change? On my boards, the CS4270 driver is compiled in-kernel, not as a module, and so when the kernel boots, the driver announces itself when it's loaded. I like this a lot, which is why I always put a
pr_info()in the__initfunction of my modules.
It's rather chatty, especially where one has unused modules compiled into the kernel, and displayed on the (usually slow serial) console by default. Having lots of drivers announce themselves in this fashion gets rather spammy, at the minute only a very small proportion do so. As a matter of policy Linux drivers generally don't do this.
Printing a message when the driver is actually binding to hardware can be useful, mainly if it's something like announcing the device revision, but doing so at module load isn't really conveying anything meaningful.
Quoted 6 lines · Mark Brown
Signed-off-by Mark Brown <[email protected]>--- sound/soc/codecs/ak4104.c | 1 - sound/soc/codecs/cs4270.c | 2 -- 2 files changed, 0 insertions(+), 3 deletions(-)