powerpc sound, some more patches

The IRQ rework caused some hiccups here, in some cases we call get_irq without a device node. This patch makes it catch that case and return NO_IRQ when it happens, along with changing the place where the irq is checked to check for NO_IRQ instead of -1.

@@ -112,7 +112,10 @@ static struct device_node *get_gpio(char
 
 static void get_irq(struct device_node * np, int *irqptr)
 {
-	*irqptr = irq_of_parse_and_map(np, 0);
+	if (np)
+		*irqptr = irq_of_parse_and_map(np, 0);
+	else
+		*irqptr = NO_IRQ;
 }
 
 /* 0x4 is outenable, 0x1 is out, thus 4 or 5 */

@@ -322,7 +325,7 @@ static int ftr_set_notify(struct gpio_ru
 		return -EINVAL;
 	}
 
-	if (irq == -1)
+	if (irq == NO_IRQ)
 		return -ENODEV;
 
 	mutex_lock(&notif->mutex);

--

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

This patch fixes the toonie codec to be actually usable.

@@ -51,6 +51,13 @@ static struct transfer_info toonie_trans
 	{}
 };
 
+static int toonie_usable(struct codec_info_item *cii,
+			 struct transfer_info *ti,
+			 struct transfer_info *out)
+{
+	return 1;
+}
+
 #ifdef CONFIG_PM
 static int toonie_suspend(struct codec_info_item *cii, pm_message_t state)
 {

@@ -69,6 +76,7 @@ static struct codec_info toonie_codec_in
 	.sysclock_factor = 256,
 	.bus_factor = 64,
 	.owner = THIS_MODULE,
+	.usable = toonie_usable,
 #ifdef CONFIG_PM
 	.suspend = toonie_suspend,
 	.resume = toonie_resume,

@@ -79,19 +87,20 @@ static int toonie_init_codec(struct aoa_
 {
 	struct toonie *toonie = codec_to_toonie(codec);
 
+	/* nothing connected? what a joke! */
+	if (toonie->codec.connected != 1)
+		return -ENOTCONN;
+
 	if (aoa_snd_device_new(SNDRV_DEV_LOWLEVEL, toonie, &ops)) {
 		printk(KERN_ERR PFX "failed to create toonie snd device!\n");
 		return -ENODEV;
 	}
 
-	/* nothing connected? what a joke! */
-	if (toonie->codec.connected != 1)
-		return -ENOTCONN;
-
 	if (toonie->codec.soundbus_dev->attach_codec(toonie->codec.soundbus_dev,
 						     aoa_get_card(),
 						     &toonie_codec_info, toonie)) {
 		printk(KERN_ERR PFX "error creating toonie pcm\n");
+		snd_device_free(aoa_get_card(), toonie);
 		return -ENODEV;
 	}
 

--

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

This patch makes snd-powermac load when it can't bind the device right away. That's the expected behaviour for hotplugging, but fixes an important problem I was seeing with doing a modprobe snd-powermac with a version that refuses loading on machines with layout-id: snd-powermac would create a bunch of uevents and then refuse to load, the uevents causing udev to reload it again, ad eternum.

@@ -181,21 +181,14 @@ static int __init alsa_card_pmac_init(vo
 	if ((err = platform_driver_register(&snd_pmac_driver)) < 0)
 		return err;
 	device = platform_device_register_simple(SND_PMAC_DRIVER, -1, NULL, 0);
-	if (!IS_ERR(device)) {
-		if (platform_get_drvdata(device))
-			return 0;
-		platform_device_unregister(device);
-		err = -ENODEV;
-	} else
-		err = PTR_ERR(device);
-	platform_driver_unregister(&snd_pmac_driver);
-	return err;
+	return 0;
 
 }
 
 static void __exit alsa_card_pmac_exit(void)
 {
-	platform_device_unregister(device);
+	if (!IS_ERR(device))
+		platform_device_unregister(device);
 	platform_driver_unregister(&snd_pmac_driver);
 }
 

--

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

This adds me into the MAINTAINERS file for the AOA driver.

@@ -292,6 +292,13 @@ L:	[email protected]
 W:	http://www.amd.com/us-en/ConnectivitySolutions/TechnicalResources/0,,50_2334_2452_11363,00.html
 S:	Supported
 
+AOA (Apple Onboard Audio) ALSA DRIVER
+P:	Johannes Berg
+M:	[email protected]
+L:	[email protected]
+L:	[email protected]
+S:	Maintained
+
 APM DRIVER
 P:	Stephen Rothwell
 M:	[email protected]

--

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

Sometimes we simply want to turn off or on everything, and when recently a warning was added when a certain platform function can't be called, this triggered all the time in those cases. This patch shows the warning only if the error was different from the function not existing.

The alternative would be to not even try calling the function when it doesn't exist by first checking which exist and then only calling those that do, but that adds complexity that isn't necessary.

@@ -18,7 +18,7 @@ static void pmf_gpio_set_##name(struct g
 							\
 	if (unlikely(!rt)) return;				\
 	rc = pmf_call_function(rt->node, #name "-mute", &args);	\
-	if (rc)							\
+	if (rc && rc != -ENODEV)				\
 		printk(KERN_WARNING "pmf_gpio_set_" #name	\
 		" failed, rc: %d\n", rc);			\
 	rt->implementation_private &= ~(1<<bit);		\

--

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

This adds me into the MAINTAINERS file for the AOA driver.

@@ -292,6 +292,13 @@ L:	[email protected]
 W:	http://www.amd.com/us-en/ConnectivitySolutions/TechnicalResources/0,,50_2334_2452_11363,00.html
 S:	Supported
 
+AOA (Apple Onboard Audio) ALSA DRIVER
+P:	Johannes Berg
+M:	[email protected]
+L:	[email protected]
+L:	[email protected]
+S:	Maintained
+
 APM DRIVER
 P:	Stephen Rothwell
 M:	[email protected]

--

Here's a new and hopefully final round of patches fixing problems with my new snd-aoa driver. There are fixes for the PPC Mac Mini as well as for a problem with snd-powermac that I had noticed.

The IRQ rework caused some hiccups here, in some cases we call get_irq without a device node. This patch makes it catch that case and return NO_IRQ when it happens, along with changing the place where the irq is checked to check for NO_IRQ instead of -1.

@@ -112,7 +112,10 @@ static struct device_node *get_gpio(char
 
 static void get_irq(struct device_node * np, int *irqptr)
 {
-	*irqptr = irq_of_parse_and_map(np, 0);
+	if (np)
+		*irqptr = irq_of_parse_and_map(np, 0);
+	else
+		*irqptr = NO_IRQ;
 }
 
 /* 0x4 is outenable, 0x1 is out, thus 4 or 5 */

@@ -322,7 +325,7 @@ static int ftr_set_notify(struct gpio_ru
 		return -EINVAL;
 	}
 
-	if (irq == -1)
+	if (irq == NO_IRQ)
 		return -ENODEV;
 
 	mutex_lock(&notif->mutex);

--

This patch fixes the toonie codec to be actually usable.

@@ -51,6 +51,13 @@ static struct transfer_info toonie_trans
 	{}
 };
 
+static int toonie_usable(struct codec_info_item *cii,
+			 struct transfer_info *ti,
+			 struct transfer_info *out)
+{
+	return 1;
+}
+
 #ifdef CONFIG_PM
 static int toonie_suspend(struct codec_info_item *cii, pm_message_t state)
 {

@@ -69,6 +76,7 @@ static struct codec_info toonie_codec_in
 	.sysclock_factor = 256,
 	.bus_factor = 64,
 	.owner = THIS_MODULE,
+	.usable = toonie_usable,
 #ifdef CONFIG_PM
 	.suspend = toonie_suspend,
 	.resume = toonie_resume,

@@ -79,19 +87,20 @@ static int toonie_init_codec(struct aoa_
 {
 	struct toonie *toonie = codec_to_toonie(codec);
 
+	/* nothing connected? what a joke! */
+	if (toonie->codec.connected != 1)
+		return -ENOTCONN;
+
 	if (aoa_snd_device_new(SNDRV_DEV_LOWLEVEL, toonie, &ops)) {
 		printk(KERN_ERR PFX "failed to create toonie snd device!\n");
 		return -ENODEV;
 	}
 
-	/* nothing connected? what a joke! */
-	if (toonie->codec.connected != 1)
-		return -ENOTCONN;
-
 	if (toonie->codec.soundbus_dev->attach_codec(toonie->codec.soundbus_dev,
 						     aoa_get_card(),
 						     &toonie_codec_info, toonie)) {
 		printk(KERN_ERR PFX "error creating toonie pcm\n");
+		snd_device_free(aoa_get_card(), toonie);
 		return -ENODEV;
 	}
 

--

This patch makes snd-powermac load when it can't bind the device right away. That's the expected behaviour for hotplugging, but fixes an important problem I was seeing with doing a modprobe snd-powermac with a version that refuses loading on machines with layout-id: snd-powermac would create a bunch of uevents and then refuse to load, the uevents causing udev to reload it again, ad eternum.

@@ -181,21 +181,14 @@ static int __init alsa_card_pmac_init(vo
 	if ((err = platform_driver_register(&snd_pmac_driver)) < 0)
 		return err;
 	device = platform_device_register_simple(SND_PMAC_DRIVER, -1, NULL, 0);
-	if (!IS_ERR(device)) {
-		if (platform_get_drvdata(device))
-			return 0;
-		platform_device_unregister(device);
-		err = -ENODEV;
-	} else
-		err = PTR_ERR(device);
-	platform_driver_unregister(&snd_pmac_driver);
-	return err;
+	return 0;
 
 }
 
 static void __exit alsa_card_pmac_exit(void)
 {
-	platform_device_unregister(device);
+	if (!IS_ERR(device))
+		platform_device_unregister(device);
 	platform_driver_unregister(&snd_pmac_driver);
 }
 

--

Sometimes we simply want to turn off or on everything, and when recently a warning was added when a certain platform function can't be called, this triggered all the time in those cases. This patch shows the warning only if the error was different from the function not existing.

The alternative would be to not even try calling the function when it doesn't exist by first checking which exist and then only calling those that do, but that adds complexity that isn't necessary.

@@ -18,7 +18,7 @@ static void pmf_gpio_set_##name(struct g
 							\
 	if (unlikely(!rt)) return;				\
 	rc = pmf_call_function(rt->node, #name "-mute", &args);	\
-	if (rc)							\
+	if (rc && rc != -ENODEV)				\
 		printk(KERN_WARNING "pmf_gpio_set_" #name	\
 		" failed, rc: %d\n", rc);			\
 	rt->implementation_private &= ~(1<<bit);		\

--

At Tue, 18 Jul 2006 19:28:41 +0200,

Thanks, now all patches are on ALSA tree.

Takashi

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

Thanks.
These should still go into 2.6.18, will that be possible?

johannes

At Tue, 25 Jul 2006 17:13:19 +0200 (CEST),

Agreed.

Jaroslav, could you update alsa git tree for pushing? We have ca. 20 pending fixes including these ppc-patches.

Takashi

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

They should still, for sure. Johannes, please sync with Paulus since I think you posted previous version of at least some of these that he's about to push to Linus via the powerpc tree.

Ben.

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

They should still, for sure. Johannes, please sync with Paulus since I think you posted previous version of at least some of these that he's about to push to Linus via the powerpc tree.

Ben.

Hm, I don't think I ever posted 'final' versions of these that I signed off, but I'll take a look.

johannes

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

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

Pretty please? People are hitting these bugs every week and I always get mail about it (and probably more people hitting the bugs who don't manage to mail me...).

johannes