Skip to contentSkip to conversation

ARM: OMAP: fix compilation problems with v3.3-rc1

PATCH New Paul Walmsley 5 participants 11 messages Jan 22, 2012

Subsystems (7)
Arm (32-bit) Arm sub-architectures I2C subsystem host drivers OMAP2+ support +3 more subsystems
lore
Share & tools
lore

Paul Walmsley

author maintainer
#0 lore raw

cover ARM: OMAP: fix compilation problems with v3.3-rc1

Fix several compilation problems detected with a 2420-only kernel configuration. This series is intended for the v3.3-rc merge window.

These patches are available via git at git://git.pwsan.com/linux-2.6 in the branch "omap2420_compilation_fixes_3.3rc"

- Paul

---

Paul Walmsley (3):
ARM: OMAP: PRM: fix missing plat/irqs.h build breakage
ARM: OMAP2+: io: fix compilation breakage on 2420-only configs
ARM: OMAP: I2C: fix compilation when !CONFIG_OF

 arch/arm/mach-omap2/io.c           |    4 +++-
 arch/arm/mach-omap2/prm2xxx_3xxx.c |    1 +
 drivers/i2c/busses/i2c-omap.c      |    4 +++-
 3 files changed, 7 insertions(+), 2 deletions(-)
author maintainer
#1 lore raw

PATCH 1/3 ARM: OMAP: PRM: fix missing plat/irqs.h build breakage

Commit 22f51371f8c35869ed850f46aa76b6cc2b502110 ("ARM: OMAP3: pm: use prcm chain handler") breaks the build on a 2420-only config, due to a missing include for plat/irqs.h:

CC arch/arm/mach-omap2/prm2xxx_3xxx.o
arch/arm/mach-omap2/prm2xxx_3xxx.c:41:11: error: 'INT_34XX_PRCM_MPU_IRQ' undeclared here (not in a function)

Fix by explicitly including it.

Signed-off-by Paul Walmsley <[email protected]>
Cc Tero Kristo <[email protected]>
Cc Kevin Hilman <[email protected]>
---
 arch/arm/mach-omap2/prm2xxx_3xxx.c |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)
arch/arm/mach-omap2/prm2xxx_3xxx.c +1 -0
index c1c4d86..9ce7654 100644
@@ -19,6 +19,7 @@
 #include "common.h"
 #include <plat/cpu.h>
 #include <plat/prcm.h>
+#include <plat/irqs.h>
 
 #include "vp.h"
author maintainer
#2 lore raw

PATCH 2/3 ARM: OMAP2+: io: fix compilation breakage on 2420-only configs

In reply to Paul Walmsley

Commit 7b250aff1ce346b6c7bc0329a2350334d1c66525 ("ARM: OMAP: Avoid cpu_is_omapxxxx usage until map_io is done") breaks the build on a 2420-only config on v3.3-rc1:

arch/arm/mach-omap2/built-in.o: In function `omap2430_init_early':
arch/arm/mach-omap2/io.c:406: undefined reference to omap2_set_globals_243x' arch/arm/mach-omap2/io.c:410: undefined reference to omap243x_clockdomains_init' arch/arm/mach-omap2/io.c:411: undefined reference to `omap2430_hwmod_init'

Fix by only compiling omap2420_init_early() when CONFIG_SOC_OMAP2420 is selected, and only compiling omap2430_init_early() when CONFIG_SOC_OMAP2430 is selected.

Signed-off-by Paul Walmsley <[email protected]>
Cc Tony Lindgren <[email protected]>
---
 arch/arm/mach-omap2/io.c |    4 +++-
 1 files changed, 3 insertions(+), 1 deletions(-)
arch/arm/mach-omap2/io.c +3 -1
index 3f174d5..eb50c29 100644
@@ -388,7 +388,7 @@ static void __init omap_hwmod_init_postsetup(void)
 	omap_pm_if_early_init();
 }
 
-#ifdef CONFIG_ARCH_OMAP2
+#ifdef CONFIG_SOC_OMAP2420
 void __init omap2420_init_early(void)
 {
 	omap2_set_globals_242x();

@@ -400,7 +400,9 @@ void __init omap2420_init_early(void)
 	omap_hwmod_init_postsetup();
 	omap2420_clk_init();
 }
+#endif
 
+#ifdef CONFIG_SOC_OMAP2430
 void __init omap2430_init_early(void)
 {
 	omap2_set_globals_243x();
author maintainer
#3 lore raw

PATCH 3/3 ARM: OMAP: I2C: fix compilation when !CONFIG_OF

In reply to Paul Walmsley

Commit 6145197be6cc0583fa1a2f4ec1079d366137061e ("i2c: OMAP: Add DT support for i2c controller") breaks compilation when CONFIG_OF is not defined:

CC drivers/i2c/busses/i2c-omap.o
drivers/i2c/busses/i2c-omap.c: In function 'omap_i2c_probe':
drivers/i2c/busses/i2c-omap.c:1021:26: error: 'omap_i2c_of_match' undeclared (first use in this function) drivers/i2c/busses/i2c-omap.c:1021:26: note: each undeclared identifier is reported only once for each function it appears in

Fix this by avoiding of_*() functions when !CONFIG_OF.

Signed-off-by Paul Walmsley <[email protected]>
Cc Benoit Cousson <[email protected]>
Cc Rob Herring <[email protected]>
Cc Grant Likely <[email protected]>
Cc Kevin Hilman <[email protected]>
Cc Tony Lindgren <[email protected]>
---
 drivers/i2c/busses/i2c-omap.c |    4 +++-
 1 files changed, 3 insertions(+), 1 deletions(-)
drivers/i2c/busses/i2c-omap.c +3 -1
index f713eac..1008e96 100644
@@ -989,7 +989,7 @@ omap_i2c_probe(struct platform_device *pdev)
 	struct resource		*mem, *irq, *ioarea;
 	struct omap_i2c_bus_platform_data *pdata = pdev->dev.platform_data;
 	struct device_node	*node = pdev->dev.of_node;
-	const struct of_device_id *match;
+	const struct of_device_id *match = NULL;
 	irq_handler_t isr;
 	int r;
 

@@ -1018,7 +1018,9 @@ omap_i2c_probe(struct platform_device *pdev)
 		goto err_release_region;
 	}
 
+#ifdef CONFIG_OF
 	match = of_match_device(omap_i2c_of_match, &pdev->dev);
+#endif
 	if (match) {
 		u32 freq = 100000; /* default to 100000 Hz */
#4 lore raw

PATCH 3/3 ARM: OMAP: I2C: fix compilation when !CONFIG_OF

Quoted 35 lines · Paul Walmsley

Commit 6145197be6cc0583fa1a2f4ec1079d366137061e ("i2c: OMAP: Add DT support for i2c controller") breaks compilation when CONFIG_OF is not defined:

CC drivers/i2c/busses/i2c-omap.o
drivers/i2c/busses/i2c-omap.c: In function 'omap_i2c_probe':
drivers/i2c/busses/i2c-omap.c:1021:26: error: 'omap_i2c_of_match' undeclared (first use in this function) drivers/i2c/busses/i2c-omap.c:1021:26: note: each undeclared identifier is reported only once for each function it appears in

Fix this by avoiding of_*() functions when !CONFIG_OF.

Signed-off-by Paul Walmsley <[email protected]>
Cc Benoit Cousson <[email protected]>
Cc Rob Herring <[email protected]>
Cc Grant Likely <[email protected]>
Cc Kevin Hilman <[email protected]>
Cc Tony Lindgren <[email protected]>
---
 drivers/i2c/busses/i2c-omap.c |    4 +++-
 1 files changed, 3 insertions(+), 1 deletions(-)
drivers/i2c/busses/i2c-omap.c +1 -1
index f713eac..1008e96 100644
@@ -989,7 +989,7 @@ omap_i2c_probe(struct platform_device *pdev)
 	struct resource		*mem, *irq, *ioarea;
 	struct omap_i2c_bus_platform_data *pdata = pdev->dev.platform_data;
 	struct device_node	*node = pdev->dev.of_node;
-	const struct of_device_id *match;
+	const struct of_device_id *match = NULL;
 	irq_handler_t isr;
 	int r;

drivers/i2c/busses/i2c-omap.c omap_i2c_probe(struct platform_device *pdev) · line 1023

Quoted 6 lines
drivers/i2c/busses/i2c-omap.c +2 -0
@@ -1018,7 +1018,9 @@ omap_i2c_probe(struct platform_device *pdev)
 		goto err_release_region;
 	}
 
+#ifdef CONFIG_OF
 	match = of_match_device(omap_i2c_of_match, &pdev->dev);
+#endif

How about an empty function for of_match_device instead to avoid the ifdef.

Rob

Quoted 10 lines
drivers/i2c/busses/i2c-omap.c +0 -0
 	if (match) {
 		u32 freq = 100000; /* default to 100000 Hz */

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#5 lore raw

PATCH 3/3 ARM: OMAP: I2C: fix compilation when !CONFIG_OF

Quoted 39 lines · Rob Herring
Quoted 38 lines · Paul Walmsley

Commit 6145197be6cc0583fa1a2f4ec1079d366137061e ("i2c: OMAP: Add DT support for i2c controller") breaks compilation when CONFIG_OF is not defined:

? CC ? ? ?drivers/i2c/busses/i2c-omap.o
drivers/i2c/busses/i2c-omap.c: In function 'omap_i2c_probe':
drivers/i2c/busses/i2c-omap.c:1021:26: error: 'omap_i2c_of_match' undeclared (first use in this function) drivers/i2c/busses/i2c-omap.c:1021:26: note: each undeclared identifier is reported only once for each function it appears in

Fix this by avoiding of_*() functions when !CONFIG_OF.

Signed-off-by Paul Walmsley <[email protected]>
Cc Benoit Cousson <[email protected]>
Cc Rob Herring <[email protected]>
Cc Grant Likely <[email protected]>
Cc Kevin Hilman <[email protected]>
Cc Tony Lindgren <[email protected]>

---
?drivers/i2c/busses/i2c-omap.c | ? ?4 +++-
?1 files changed, 3 insertions(+), 1 deletions(-)

drivers/i2c/busses/i2c-omap.c +1 -1
index f713eac..1008e96 100644
@@ -989,7 +989,7 @@ omap_i2c_probe(struct platform_device *pdev)
 ? ? ? struct resource ? ? ? ? *mem, *irq, *ioarea;
 ? ? ? struct omap_i2c_bus_platform_data *pdata = pdev->dev.platform_data;
 ? ? ? struct device_node ? ? ?*node = pdev->dev.of_node;
- ? ? const struct of_device_id *match;
+ ? ? const struct of_device_id *match = NULL;
 ? ? ? irq_handler_t isr;
 ? ? ? int r;
 

@@ -1018,7 +1018,9 @@ omap_i2c_probe(struct platform_device *pdev)
 ? ? ? ? ? ? ? goto err_release_region;
 ? ? ? }
Quoted 5 lines

drivers/i2c/busses/i2c-omap.c omap_i2c_probe(struct platform_device *pdev) · line 1023

Quoted 3 lines
drivers/i2c/busses/i2c-omap.c +2 -0
+#ifdef CONFIG_OF
 ? ? ? match = of_match_device(omap_i2c_of_match, &pdev->dev);
+#endif

How about an empty function for of_match_device instead to avoid the ifdef.

That already exists. The problem is the undefined omap_i2c_of_match table reference.

g.

Signature · 2 lines
Grant Likely, B.Sc., P.Eng.
Secret Lab Technologies Ltd.
#6 lore raw

PATCH 3/3 ARM: OMAP: I2C: fix compilation when !CONFIG_OF

In reply to Paul Walmsley
Quoted 10 lines · Paul Walmsley

Commit 6145197be6cc0583fa1a2f4ec1079d366137061e ("i2c: OMAP: Add DT support for i2c controller") breaks compilation when CONFIG_OF is not defined:

CC drivers/i2c/busses/i2c-omap.o
drivers/i2c/busses/i2c-omap.c: In function 'omap_i2c_probe':
drivers/i2c/busses/i2c-omap.c:1021:26: error: 'omap_i2c_of_match' undeclared (first use in this function) drivers/i2c/busses/i2c-omap.c:1021:26: note: each undeclared identifier is reported only once for each function it appears in

Fix this by avoiding of_*() functions when !CONFIG_OF.

To avoid #ifdef blocks, how about one of these fixes instead? of_match_device resolves to a static inline NULL when CONFIG_OF is not selected.

g.

---

drivers/i2c/busses/i2c-omap.c +2 -0
index f713eac..c22e51f 100644
@@ -979,6 +979,8 @@ static const struct of_device_id omap_i2c_of_match[] = {
 	{ },
 };
 MODULE_DEVICE_TABLE(of, omap_i2c_of_match);
+#else
+#define omap_i2c_of_match NULL
 #endif
 
 static int __devinit

Or alternately:
---

drivers/i2c/busses/i2c-omap.c +1 -1
index f713eac..801df60 100644
@@ -1018,7 +1018,7 @@ omap_i2c_probe(struct platform_device *pdev)
 		goto err_release_region;
 	}
 
-	match = of_match_device(omap_i2c_of_match, &pdev->dev);
+	match = of_match_device(of_match_ptr(omap_i2c_of_match), &pdev->dev);
 	if (match) {
 		u32 freq = 100000; /* default to 100000 Hz */
#7 lore raw

PATCH 3/3 ARM: OMAP: I2C: fix compilation when !CONFIG_OF

Quoted 24 lines · Grant Likely
Quoted 10 lines · Paul Walmsley

Commit 6145197be6cc0583fa1a2f4ec1079d366137061e ("i2c: OMAP: Add DT support for i2c controller") breaks compilation when CONFIG_OF is not defined:

CC drivers/i2c/busses/i2c-omap.o
drivers/i2c/busses/i2c-omap.c: In function 'omap_i2c_probe':
drivers/i2c/busses/i2c-omap.c:1021:26: error: 'omap_i2c_of_match' undeclared (first use in this function) drivers/i2c/busses/i2c-omap.c:1021:26: note: each undeclared identifier is reported only once for each function it appears in

Fix this by avoiding of_*() functions when !CONFIG_OF.

To avoid #ifdef blocks, how about one of these fixes instead? of_match_device resolves to a static inline NULL when CONFIG_OF is not selected.

g.

---

drivers/i2c/busses/i2c-omap.c +0 -0
index f713eac..c22e51f 100644

drivers/i2c/busses/i2c-omap.c static const struct of_device_id omap_i2c_of_match[] = { · line 984

Quoted 6 lines
drivers/i2c/busses/i2c-omap.c +2 -0
@@ -979,6 +979,8 @@ static const struct of_device_id omap_i2c_of_match[] = {
 	{ },
 };
 MODULE_DEVICE_TABLE(of, omap_i2c_of_match);
+#else
+#define omap_i2c_of_match NULL
 #endif

Or just remove the existing ifdef. Isn't OMAP going to be completely converted to DT?

I think building CONFIG_OF and !CONFIG_OF and running non-DT machines with OF enabled is going to be a constant source of breakage. We should move selecting CONFIG_OF up from boards/SOC config to family config (i.e. all OMAP).

Rob

#8 lore raw

PATCH 3/3 ARM: OMAP: I2C: fix compilation when !CONFIG_OF

Quoted 35 lines · Rob Herring
Quoted 24 lines · Grant Likely
Quoted 10 lines · Paul Walmsley

Commit 6145197be6cc0583fa1a2f4ec1079d366137061e ("i2c: OMAP: Add DT support for i2c controller") breaks compilation when CONFIG_OF is not defined:

CC drivers/i2c/busses/i2c-omap.o
drivers/i2c/busses/i2c-omap.c: In function 'omap_i2c_probe':
drivers/i2c/busses/i2c-omap.c:1021:26: error: 'omap_i2c_of_match' undeclared (first use in this function) drivers/i2c/busses/i2c-omap.c:1021:26: note: each undeclared identifier is reported only once for each function it appears in

Fix this by avoiding of_*() functions when !CONFIG_OF.

To avoid #ifdef blocks, how about one of these fixes instead? of_match_device resolves to a static inline NULL when CONFIG_OF is not selected.

g.

---

drivers/i2c/busses/i2c-omap.c +0 -0
index f713eac..c22e51f 100644

drivers/i2c/busses/i2c-omap.c static const struct of_device_id omap_i2c_of_match[] = { · line 984

Quoted 6 lines
drivers/i2c/busses/i2c-omap.c +2 -0
@@ -979,6 +979,8 @@ static const struct of_device_id omap_i2c_of_match[] = {
  	{ },
  };
  MODULE_DEVICE_TABLE(of, omap_i2c_of_match);
+#else
+#define omap_i2c_of_match NULL
  #endif

Or just remove the existing ifdef. Isn't OMAP going to be completely converted to DT?

Quoted 4 lines

I think building CONFIG_OF and !CONFIG_OF and running non-DT machines with OF enabled is going to be a constant source of breakage. We should move selecting CONFIG_OF up from boards/SOC config to family config (i.e. all OMAP).

Yep, I do agree, and even sent a patch for that.

http://lists.ozlabs.org/pipermail/devicetree-discuss/2011-December/010541.html

But it looks like it was stuck in linux-omap dt branch and not pulled for 3.3. Maybe Tony can try to sent it during -rc phase.

That being said, this will be valid for OMAP2+ only and will still break OMAP1 platform :-(

Tony reported that 2 days ago, and I sent the fix for it.
http://www.mail-archive.com/linux-omap at vger.kernel.org/msg61520.html

It is now waiting for Ben's Acked-by.

Regards,
Benoit

author maintainer
#9 lore raw

PATCH 3/3 ARM: OMAP: I2C: fix compilation when !CONFIG_OF

cc Ben, Jean

Quoted 10 lines · Benoit

http://lists.ozlabs.org/pipermail/devicetree-discuss/2011-December/010541.html

But it looks like it was stuck in linux-omap dt branch and not pulled for 3.3. Maybe Tony can try to sent it during -rc phase.

That being said, this will be valid for OMAP2+ only and will still break OMAP1 platform :-(

Tony reported that 2 days ago, and I sent the fix for it.
http://www.mail-archive.com/linux-omap at vger.kernel.org/msg61520.html

Will drop patch 3 from my series. Incidentally, it doesn't seem necessary to wait for Ben's ack after commit 046d0a37024a5b1e66ec9314e5e7feba42feb811 ("ARM: OMAP: I2C: fix compilation when !CONFIG_OF"), unless Ben disagrees with that change.

- Paul

#10 lore raw

PATCH 3/3 ARM: OMAP: I2C: fix compilation when !CONFIG_OF

Quoted 1 line · Paul Walmsley

cc Ben, Jean

I'm not involved in this at all.

Jean

Quoted 21 lines · Benoit
Quoted 11 lines

http://lists.ozlabs.org/pipermail/devicetree-discuss/2011-December/010541.html

But it looks like it was stuck in linux-omap dt branch and not pulled for 3.3. Maybe Tony can try to sent it during -rc phase.

That being said, this will be valid for OMAP2+ only and will still break OMAP1 platform :-(

Tony reported that 2 days ago, and I sent the fix for it.
http://www.mail-archive.com/linux-omap at vger.kernel.org/msg61520.html

Will drop patch 3 from my series. Incidentally, it doesn't seem necessary to wait for Ben's ack after commit 046d0a37024a5b1e66ec9314e5e7feba42feb811 ("ARM: OMAP: I2C: fix compilation when !CONFIG_OF"), unless Ben disagrees with that change.

- Paul