Re: N800 problems with MMC, LM8323 on current linux-omap git head

Re: N800 problems with MMC, LM8323 on current linux-omap git head

[Riku Voipio]

[-- Attachment #1: Type: text/plain, Size: 524 bytes --]

> Subject: [PATCH] input: add more error checks to lm8323 driver

> If we can't reach the driver, we stop trying to probe
> it. Useful when building kernel for n800 and n810.

> n800 doesn't have lm8323, so that driver shouldn't probe
> there.

Hmm. I think it would be better to avoid calling the probe
in first place. git diff made the attached patch look more
confusing than it actually is. Personally I'd eliminate
all ifdefs from those i2c_board_info structs.


-- 
"rm -rf" only sounds scary if you don't have backups

[-- Attachment #2: 0001-Separate-i2c_board_info-for-n800-and-n810.patch --]
[-- Type: text/plain, Size: 1988 bytes --]

>From 574835f1532f0be01e0456f98bf13ebb954bc6c3 Mon Sep 17 00:00:00 2001
From: Riku Voipio <riku.voipio@iki.fi>
Date: Tue, 22 Jul 2008 00:29:31 +0300
Subject: [PATCH] Separate i2c_board_info for n800 and n810

n800 and n810 have different peripherals, on the second i2c bus
(tea5761 on n800 and lm8323 on n810). Thus they need different
i2c_board_info as well to avoid probing nonexistent devices.
---
 arch/arm/mach-omap2/board-n800.c |   24 +++++++++++++++++++++---
 1 files changed, 21 insertions(+), 3 deletions(-)

diff --git a/arch/arm/mach-omap2/board-n800.c b/arch/arm/mach-omap2/board-n800.c
index ae85c2c..95a2727 100644
--- a/arch/arm/mach-omap2/board-n800.c
+++ b/arch/arm/mach-omap2/board-n800.c
@@ -654,13 +654,26 @@ static struct i2c_board_info __initdata_or_module n800_i2c_board_info_2[] = {
 		I2C_BOARD_INFO("tea5761", 0x10),
 	},
 #endif
-#ifdef CONFIG_MACH_NOKIA_N810
+	{
+		I2C_BOARD_INFO("tsl2563", 0x29),
+	},
+	{
+		I2C_BOARD_INFO("lp5521", 0x32),
+	},
+};
+
+static struct i2c_board_info __initdata_or_module n810_i2c_board_info_2[] = {
+#if defined (CONFIG_VIDEO_TCM825X) || defined (CONFIG_VIDEO_TCM825X_MODULE)
+	{
+		I2C_BOARD_INFO(TCM825X_NAME, TCM825X_I2C_ADDR),
+		.platform_data = &n800_tcm825x_platform_data,
+	},
+#endif
 	{
 		I2C_BOARD_INFO("lm8323", 0x45),
 		.irq		= OMAP_GPIO_IRQ(109),
 		.platform_data	= &lm8323_pdata,
 	},
-#endif
 	{
 		I2C_BOARD_INFO("tsl2563", 0x29),
 	},
@@ -690,8 +703,13 @@ void __init nokia_n800_common_init(void)
 	omap_serial_init();
 	omap_register_i2c_bus(1, 400, n800_i2c_board_info_1,
 			      ARRAY_SIZE(n800_i2c_board_info_1));
-	omap_register_i2c_bus(2, 400, n800_i2c_board_info_2,
+	if (machine_is_nokia_n800())
+		omap_register_i2c_bus(2, 400, n800_i2c_board_info_2,
 			      ARRAY_SIZE(n800_i2c_board_info_2));
+	if (machine_is_nokia_n810())
+		omap_register_i2c_bus(2, 400, n810_i2c_board_info_2,
+			      ARRAY_SIZE(n810_i2c_board_info_2));
+		
 	mipid_dev_init();
 	blizzard_dev_init();
 }
-- 
1.5.5.4

Re: N800 problems with MMC, LM8323 on current linux-omap git head

[Felipe Balbi]

On Tue, 22 Jul 2008 15:47:43 +0300, Riku Voipio <riku.voipio@iki.fi> wrote:

> Hmm. I think it would be better to avoid calling the probe
> in first place. git diff made the attached patch look more
> confusing than it actually is. Personally I'd eliminate
> all ifdefs from those i2c_board_info structs.

I did the same thing before, but then we're gonna start adding
several i2c_board_info for different hw. I mean, we should
just reuse the code and the driver should know when the device
is not present and stop probing without any problems to the rest
of the system.

-- 
Best Regards,

Felipe Balbi
http://blog.felipebalbi.com
me@felipebalbi.com

Re: N800 problems with MMC, LM8323 on current linux-omap git head

[Riku Voipio]

On Tue, Jul 22, 2008 at 07:52:44AM -0500, Felipe Balbi wrote:
> On Tue, 22 Jul 2008 15:47:43 +0300, Riku Voipio <riku.voipio@iki.fi> wrote:
> 
> > Hmm. I think it would be better to avoid calling the probe
> > in first place. git diff made the attached patch look more
> > confusing than it actually is. Personally I'd eliminate
> > all ifdefs from those i2c_board_info structs.

> I did the same thing before, but then we're gonna start adding
> several i2c_board_info for different hw. I mean, we should
> just reuse the code and the driver should know when the device
> is not present and stop probing without any problems to the rest
> of the system.

If the different hw has different devices on the i2c bus, then
different i2c_board_info struct should be provided. IIRC i2c_board_info
is expected to provide a accurate picture of what i2c devices is
connected to the bus. This has been used as a argument to remove detection
code from .probe functions.

To all the confused i2c list readers, the topic is about should we
split n800_i2c_board_info_2 [0] to n800 and n810 specific structs[1]
or make the lm8323 configure to err out earlier[2].

[0] http://git.kernel.org/?p=linux/kernel/git/tmlind/linux-omap-2.6.git;a=blob;f=arch/arm/mach-omap2/board-n800.c;h=ae85c2c60534820856c8bb0e019be29f2375470b;hb=HEAD#l645
[1] http://article.gmane.org/gmane.linux.ports.arm.omap/9562
[2] http://article.gmane.org/gmane.linux.ports.arm.omap/9522


"rm -rf" only sounds scary if you don't have backups

Re: [i2c] N800 problems with MMC, LM8323 on current linux-omap git head

[Ben Dooks]

On Tue, Jul 22, 2008 at 04:37:42PM +0300, Riku Voipio wrote:
> On Tue, Jul 22, 2008 at 07:52:44AM -0500, Felipe Balbi wrote:
> > On Tue, 22 Jul 2008 15:47:43 +0300, Riku Voipio <riku.voipio@iki.fi> wrote:
> > 
> > > Hmm. I think it would be better to avoid calling the probe
> > > in first place. git diff made the attached patch look more
> > > confusing than it actually is. Personally I'd eliminate
> > > all ifdefs from those i2c_board_info structs.
> 
> > I did the same thing before, but then we're gonna start adding
> > several i2c_board_info for different hw. I mean, we should
> > just reuse the code and the driver should know when the device
> > is not present and stop probing without any problems to the rest
> > of the system.
> 
> If the different hw has different devices on the i2c bus, then
> different i2c_board_info struct should be provided. IIRC i2c_board_info
> is expected to provide a accurate picture of what i2c devices is
> connected to the bus. This has been used as a argument to remove detection
> code from .probe functions.
> 
> To all the confused i2c list readers, the topic is about should we
> split n800_i2c_board_info_2 [0] to n800 and n810 specific structs[1]
> or make the lm8323 configure to err out earlier[2].

Split them into a common and then machine specfic structures, you are
allowed to register more than one board per bus.
 
> [0] http://git.kernel.org/?p=linux/kernel/git/tmlind/linux-omap-2.6.git;a=blob;f=arch/arm/mach-omap2/board-n800.c;h=ae85c2c60534820856c8bb0e019be29f2375470b;hb=HEAD#l645
> [1] http://article.gmane.org/gmane.linux.ports.arm.omap/9562
> [2] http://article.gmane.org/gmane.linux.ports.arm.omap/9522
> 
> 
> "rm -rf" only sounds scary if you don't have backups
> 
> _______________________________________________
> i2c mailing list
> i2c@lm-sensors.org
> http://lists.lm-sensors.org/mailman/listinfo/i2c

-- 
Ben (ben@fluff.org, http://www.fluff.org/)

  'a smiley only costs 4 bytes'

Re: [i2c] N800 problems with MMC, LM8323 on current linux-omap git head

[Felipe Balbi]

Hi,

On Tue, 22 Jul 2008 14:43:14 +0100, Ben Dooks <ben-linux@fluff.org> wrote:
> Split them into a common and then machine specfic structures, you are
> allowed to register more than one board per bus.

Good to know :-)

Riku, could you update your patch then ?

but it anyways doesn't annul the need of a better error handling
in lm8323. I think both patches will need to be applied :-)

-- 
Best Regards,

Felipe Balbi
http://blog.felipebalbi.com
me@felipebalbi.com

Re: [i2c] N800 problems with MMC, LM8323 on current linux-omap git head

[Riku Voipio]

[-- Attachment #1: Type: text/plain, Size: 839 bytes --]

On Tue, Jul 22, 2008 at 09:20:46AM -0500, Felipe Balbi wrote:
> On Tue, 22 Jul 2008 14:43:14 +0100, Ben Dooks <ben-linux@fluff.org> wrote:
> > Split them into a common and then machine specfic structures, you are
> > allowed to register more than one board per bus.

> Good to know :-)

Indeed - compared to my previous version, this one makes the boardfile
even smaller:

   text    data     bss     dec     hex filename
   2362    1552       4    3918     f4e old/arch/arm/mach-omap2/board-n800.o
   2354    1456       4    3814     ee6 new/arch/arm/mach-omap2/board-n800.o


> Riku, could you update your patch then ?

tested and attached.

> but it anyways doesn't annul the need of a better error handling
> in lm8323. I think both patches will need to be applied :-)

true. 

-- 
"rm -rf" only sounds scary if you don't have backups

[-- Attachment #2: 0001-Separate-i2c_board_info-for-n800-and-n810.patch --]
[-- Type: text/plain, Size: 2690 bytes --]

>From 1e9d7271505e85c99fe9b46653569ff6216e07b3 Mon Sep 17 00:00:00 2001
From: Riku Voipio <riku.voipio@iki.fi>
Date: Tue, 22 Jul 2008 12:29:31 +0300
Subject: [PATCH] Separate i2c_board_info for n800 and n810

n800 and n810 have different peripherals on the second i2c bus
(tea5761 on n800 and lm8323 on n810). Split the i2c_board_info to
common and hw specific to avoid probing nonexistent devices.

Signed-off-by: Riku Voipio <riku.voipio@iki.fi>
---
 arch/arm/mach-omap2/board-n800.c |   36 ++++++++++++++++++++++--------------
 1 files changed, 22 insertions(+), 14 deletions(-)

diff --git a/arch/arm/mach-omap2/board-n800.c b/arch/arm/mach-omap2/board-n800.c
index ae85c2c..e156dbd 100644
--- a/arch/arm/mach-omap2/board-n800.c
+++ b/arch/arm/mach-omap2/board-n800.c
@@ -642,30 +642,31 @@ static struct i2c_board_info __initdata n800_i2c_board_info_1[] = {
 
 extern struct tcm825x_platform_data n800_tcm825x_platform_data;
 
-static struct i2c_board_info __initdata_or_module n800_i2c_board_info_2[] = {
-#if defined (CONFIG_VIDEO_TCM825X) || defined (CONFIG_VIDEO_TCM825X_MODULE)
+static struct i2c_board_info __initdata_or_module n8x0_i2c_board_info_2[] = {
 	{
 		I2C_BOARD_INFO(TCM825X_NAME, TCM825X_I2C_ADDR),
 		.platform_data = &n800_tcm825x_platform_data,
 	},
-#endif
-#if defined(CONFIG_RADIO_TEA5761) || defined(CONFIG_RADIO_TEA5761_MODULE)
 	{
-		I2C_BOARD_INFO("tea5761", 0x10),
+		I2C_BOARD_INFO("tsl2563", 0x29),
 	},
-#endif
-#ifdef CONFIG_MACH_NOKIA_N810
 	{
-		I2C_BOARD_INFO("lm8323", 0x45),
-		.irq		= OMAP_GPIO_IRQ(109),
-		.platform_data	= &lm8323_pdata,
+		I2C_BOARD_INFO("lp5521", 0x32),
 	},
-#endif
+};
+
+
+static struct i2c_board_info __initdata_or_module n800_i2c_board_info_2[] = {
 	{
-		I2C_BOARD_INFO("tsl2563", 0x29),
+		I2C_BOARD_INFO("tea5761", 0x10),
 	},
+};
+
+static struct i2c_board_info __initdata_or_module n810_i2c_board_info_2[] = {
 	{
-		I2C_BOARD_INFO("lp5521", 0x32),
+		I2C_BOARD_INFO("lm8323", 0x45),
+		.irq		= OMAP_GPIO_IRQ(109),
+		.platform_data	= &lm8323_pdata,
 	},
 };
 
@@ -690,8 +691,15 @@ void __init nokia_n800_common_init(void)
 	omap_serial_init();
 	omap_register_i2c_bus(1, 400, n800_i2c_board_info_1,
 			      ARRAY_SIZE(n800_i2c_board_info_1));
-	omap_register_i2c_bus(2, 400, n800_i2c_board_info_2,
+	omap_register_i2c_bus(2, 400, n8x0_i2c_board_info_2,
 			      ARRAY_SIZE(n800_i2c_board_info_2));
+	if (machine_is_nokia_n800())
+		i2c_register_board_info(2, n800_i2c_board_info_2,
+			ARRAY_SIZE(n800_i2c_board_info_2));
+	if (machine_is_nokia_n810())
+	    i2c_register_board_info(2, n810_i2c_board_info_2,
+			ARRAY_SIZE(n810_i2c_board_info_2));
+		
 	mipid_dev_init();
 	blizzard_dev_init();
 }
-- 
1.5.5.4

Re: [i2c] N800 problems with MMC, LM8323 on current linux-omap git head

[Felipe Balbi]

On Wed, 23 Jul 2008 12:38:49 +0300, Riku Voipio <riku.voipio@iki.fi> wrote:
> Indeed - compared to my previous version, this one makes the boardfile
> even smaller:
> 
>    text    data     bss     dec     hex filename
>    2362    1552       4    3918     f4e
> old/arch/arm/mach-omap2/board-n800.o
>    2354    1456       4    3814     ee6
> new/arch/arm/mach-omap2/board-n800.o

Cool :-)

Acked-by: Felipe Balbi <felipe.balbi@nokia.com>

-- 
Best Regards,

Felipe Balbi
http://blog.felipebalbi.com
me@felipebalbi.com

Re: [i2c] N800 problems with MMC, LM8323 on current linux-omap git head

[Tony Lindgren]

* Felipe Balbi <me@felipebalbi.com> [080723 13:13]:
> 
> 
> On Wed, 23 Jul 2008 12:38:49 +0300, Riku Voipio <riku.voipio@iki.fi> wrote:
> > Indeed - compared to my previous version, this one makes the boardfile
> > even smaller:
> > 
> >    text    data     bss     dec     hex filename
> >    2362    1552       4    3918     f4e
> > old/arch/arm/mach-omap2/board-n800.o
> >    2354    1456       4    3814     ee6
> > new/arch/arm/mach-omap2/board-n800.o
> 
> Cool :-)

Pushing to linux-omap tree.

Tony


> Acked-by: Felipe Balbi <felipe.balbi@nokia.com>
> 
> -- 
> Best Regards,
> 
> Felipe Balbi
> http://blog.felipebalbi.com
> me@felipebalbi.com
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-omap" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html