smc91x section fix

PATCH In review David Brownell 3 participants 7 messages Feb 25, 2008 – Jul 25, 2008

Subsystems (2)

Networking Network drivers

lore

Share & tools

lore

David Brownell

author

2008-02-25 03:34 UTC #0 lore raw

PATCH 1/1 smc91x section fix

Section fixup:

WARNING: drivers/net/built-in.o(.text+0x1a2c): Section mismatch
in reference from the function smc_drv_probe() to the function .init.text:smc_probe()
The function smc_drv_probe() references
the function __init smc_probe().
This is often because smc_drv_probe lacks a __init annotation or the annotation of smc_probe is wrong.

Signed-off-by David Brownell <[email protected]>

---
 drivers/net/smc91x.c |    2 +-
 1 files changed, 1 insertion(+), 1 deletion(-)

drivers/net/smc91x.c +1 -1

@@ -2121,7 +2121,7 @@ static void smc_release_datacs(struct pl
  *	0 --> there is a device
  *	anything else, error
  */
-static int smc_drv_probe(struct platform_device *pdev)
+static int __init smc_drv_probe(struct platform_device *pdev)
 {
 	struct net_device *ndev;
 	struct resource *res, *ires;

Sam Ravnborg

2008-02-25 06:02 UTC #1 lore raw

Quoted 27 lines · David Brownell

Section fixup:
WARNING: drivers/net/built-in.o(.text+0x1a2c): Section mismatch
in reference from the function smc_drv_probe() to the function .init.text:smc_probe()
The function smc_drv_probe() references
the function __init smc_probe().
This is often because smc_drv_probe lacks a __init annotation or the annotation of smc_probe is wrong.
Signed-off-by David Brownell <[email protected]>
---
 drivers/net/smc91x.c |    2 +-
 1 files changed, 1 insertion(+), 1 deletion(-)
drivers/net/smc91x.c +1 -1
@@ -2121,7 +2121,7 @@ static void smc_release_datacs(struct pl
  *	0 --> there is a device
  *	anything else, error
  */
-static int smc_drv_probe(struct platform_device *pdev)
+static int __init smc_drv_probe(struct platform_device *pdev)
 {
 	struct net_device *ndev;
 	struct resource *res, *ires;

Quoted 1 line

From a quick look this is wrong.

smc_drv_probe is assined the .probe member so it is used during hotplug and thus should be __devinit.
Likewise smc_probe is used by smc_drv_probe and thus smc_probe should be __devinit too.

So something like the appended. On my way out of the door so not even build tested!

Sam

drivers/net/smc91x.c +2 -2

index 4020e9e..640bca5 100644
@@ -1775,7 +1775,7 @@ static int __init smc_findirq(void __iomem *ioaddr)
  * o  actually GRAB the irq.
  * o  GRAB the region
  */
-static int __init smc_probe(struct net_device *dev, void __iomem *ioaddr,
+static int __devinit smc_probe(struct net_device *dev, void __iomem *ioaddr,
 			    unsigned long irq_flags)
 {
 	struct smc_local *lp = netdev_priv(dev);

@@ -2121,7 +2121,7 @@ static void smc_release_datacs(struct platform_device *pdev, struct net_device *
  *	0 --> there is a device
  *	anything else, error
  */
-static int smc_drv_probe(struct platform_device *pdev)
+static int __devinit smc_drv_probe(struct platform_device *pdev)
 {
 	struct net_device *ndev;
 	struct resource *res, *ires;

David Brownell

author

2008-02-25 06:33 UTC #2 lore raw

Quoted 5 lines · Sam Ravnborg

From a quick look this is wrong.
smc_drv_probe is assined the .probe member so it is used during hotplug and thus should be __devinit.
Likewise smc_probe is used by smc_drv_probe and thus smc_probe should be __devinit too.

Thing is, with only rare exceptions, devices on the platform bus are *NOT* hotpluggable. So using __devinit/__devexit and friends adds up to no more than a waste of I-space.

Nico may know if this driver is one of the rare exceptions. Example, is it used in PCMCIA cards?

See also platform_driver_probe(), which is a much safer way to handle such issues. I've seen it save a full page in some drivers, but usually it's less.

- Dave

Sam Ravnborg

2008-02-25 20:11 UTC #3 lore raw

Quoted 11 lines · David Brownell

Quoted 5 lines · Sam Ravnborg
From a quick look this is wrong.
smc_drv_probe is assined the .probe member so it is used during hotplug and thus should be __devinit.
Likewise smc_probe is used by smc_drv_probe and thus smc_probe should be __devinit too.
Thing is, with only rare exceptions, devices on the platform bus are *NOT* hotpluggable. So using __devinit/__devexit and friends adds up to no more than a waste of I-space.

It was a quick look - thanks for the explanation.

Sam

David Brownell

author

2008-04-19 03:01 UTC #4 lore raw

PATCH 1/1 smc91x section fix

In reply to David Brownell

From: David Brownell <[email protected]>

Section fixup:

Also switch to platform_driver_probe(), and mark exit code as such. Total footprint shrinkage is about 2KB on ARMv5.

Note that few platform devices are actually hotpluggable, and these smc devices are evidently not exceptions to that rule ... else there would have been oopsing in this driver from the smc_probe() problem.

Signed-off-by David Brownell <[email protected]>

---
Changes since earlier version: use platform_driver_probe(), add the __exit annotations, not how we can know this driver doesn't need to use hotplug-supporting "__devinit" annotations.

 drivers/net/smc91x.c |   12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

drivers/net/smc91x.c +6 -6

@@ -2097,7 +2097,8 @@ static inline void smc_request_datacs(st
 	}
 }
 
-static void smc_release_datacs(struct platform_device *pdev, struct net_device *ndev)
+static void __exit smc_release_datacs(struct platform_device *pdev,
+		struct net_device *ndev)
 {
 	if (SMC_CAN_USE_DATACS) {
 		struct smc_local *lp = netdev_priv(ndev);

@@ -2124,7 +2125,7 @@ static void smc_release_datacs(struct pl
  *	0 --> there is a device
  *	anything else, error
  */
-static int smc_drv_probe(struct platform_device *pdev)
+static int __init smc_drv_probe(struct platform_device *pdev)
 {
 	struct net_device *ndev;
 	struct resource *res, *ires;

@@ -2214,7 +2215,7 @@ static int smc_drv_probe(struct platform
 	return ret;
 }
 
-static int smc_drv_remove(struct platform_device *pdev)
+static int __exit smc_drv_remove(struct platform_device *pdev)
 {
 	struct net_device *ndev = platform_get_drvdata(pdev);
 	struct smc_local *lp = netdev_priv(ndev);

@@ -2278,8 +2279,7 @@ static int smc_drv_resume(struct platfor
 }
 
 static struct platform_driver smc_driver = {
-	.probe		= smc_drv_probe,
-	.remove		= smc_drv_remove,
+	.remove		= __exit_p(smc_drv_remove),
 	.suspend	= smc_drv_suspend,
 	.resume		= smc_drv_resume,
 	.driver		= {

@@ -2299,7 +2299,7 @@ static int __init smc_init(void)
 #endif
 #endif
 
-	return platform_driver_register(&smc_driver);
+	return platform_driver_probe(&smc_driver, smc_drv_probe);
 }
 
 static void __exit smc_cleanup(void)

Jeff Garzik

Applied

2008-04-25 06:02 UTC #5 lore raw

Quoted 69 lines · David Brownell

From: David Brownell <[email protected]>
Section fixup:
WARNING: drivers/net/built-in.o(.text+0x1a2c): Section mismatch
in reference from the function smc_drv_probe() to the function .init.text:smc_probe()
The function smc_drv_probe() references
the function __init smc_probe().
This is often because smc_drv_probe lacks a __init annotation or the annotation of smc_probe is wrong.
Also switch to platform_driver_probe(), and mark exit code as such. Total footprint shrinkage is about 2KB on ARMv5.
Note that few platform devices are actually hotpluggable, and these smc devices are evidently not exceptions to that rule ... else there would have been oopsing in this driver from the smc_probe() problem.
Signed-off-by David Brownell <[email protected]>
---
Changes since earlier version: use platform_driver_probe(), add the __exit annotations, not how we can know this driver doesn't need to use hotplug-supporting "__devinit" annotations.
 drivers/net/smc91x.c |   12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)
drivers/net/smc91x.c +5 -5
@@ -2097,7 +2097,8 @@ static inline void smc_request_datacs(st
 	}
 }
 
-static void smc_release_datacs(struct platform_device *pdev, struct net_device *ndev)
+static void __exit smc_release_datacs(struct platform_device *pdev,
+		struct net_device *ndev)
 {
 	if (SMC_CAN_USE_DATACS) {
 		struct smc_local *lp = netdev_priv(ndev);
@@ -2124,7 +2125,7 @@ static void smc_release_datacs(struct pl
  *	0 --> there is a device
  *	anything else, error
  */
-static int smc_drv_probe(struct platform_device *pdev)
+static int __init smc_drv_probe(struct platform_device *pdev)
 {
 	struct net_device *ndev;
 	struct resource *res, *ires;
@@ -2214,7 +2215,7 @@ static int smc_drv_probe(struct platform
 	return ret;
 }
 
-static int smc_drv_remove(struct platform_device *pdev)
+static int __exit smc_drv_remove(struct platform_device *pdev)
 {
 	struct net_device *ndev = platform_get_drvdata(pdev);
 	struct smc_local *lp = netdev_priv(ndev);
@@ -2278,8 +2279,7 @@ static int smc_drv_resume(struct platfor
 }
 
 static struct platform_driver smc_driver = {
-	.probe		= smc_drv_probe,
-	.remove		= smc_drv_remove,
+	.remove		= __exit_p(smc_drv_remove),
 	.suspend	= smc_drv_suspend,
 	.resume		= smc_drv_resume,
 	.driver		= {

drivers/net/smc91x.c static int __init smc_init(void)

Quoted 9 lines

drivers/net/smc91x.c +1 -2

@@ -2299,7 +2299,7 @@ static int __init smc_init(void)
 #endif
 #endif
 
-	return platform_driver_register(&smc_driver);
+	return platform_driver_probe(&smc_driver, smc_drv_probe);
 }
 
 static void __exit smc_cleanup(void)
--

applied

David Brownell

author

2008-07-25 20:12 UTC #6 lore raw

In reply to David Brownell

From: David Brownell <[email protected]>

Section fixup:

Also switch to platform_driver_probe(), and mark the exit code as such. Total footprint shrinkage is about 2KB on ARMv5.

Signed-off-by David Brownell <[email protected]>

---
I thought this was queued for 2.6.26, or (later) 2.6.27 ...

 drivers/net/smc91x.c |   12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

drivers/net/smc91x.c +6 -6

@@ -2096,7 +2096,8 @@ static inline void smc_request_datacs(st
 	}
 }
 
-static void smc_release_datacs(struct platform_device *pdev, struct net_device *ndev)
+static void __exit smc_release_datacs(struct platform_device *pdev,
+		struct net_device *ndev)
 {
 	if (SMC_CAN_USE_DATACS) {
 		struct smc_local *lp = netdev_priv(ndev);

@@ -2123,7 +2124,7 @@ static void smc_release_datacs(struct pl
  *	0 --> there is a device
  *	anything else, error
  */
-static int smc_drv_probe(struct platform_device *pdev)
+static int __init smc_drv_probe(struct platform_device *pdev)
 {
 	struct smc91x_platdata *pd = pdev->dev.platform_data;
 	struct smc_local *lp;

@@ -2232,7 +2233,7 @@ static int smc_drv_probe(struct platform
 	return ret;
 }
 
-static int smc_drv_remove(struct platform_device *pdev)
+static int __exit smc_drv_remove(struct platform_device *pdev)
 {
 	struct net_device *ndev = platform_get_drvdata(pdev);
 	struct smc_local *lp = netdev_priv(ndev);

@@ -2296,8 +2297,7 @@ static int smc_drv_resume(struct platfor
 }
 
 static struct platform_driver smc_driver = {
-	.probe		= smc_drv_probe,
-	.remove		= smc_drv_remove,
+	.remove		= __exit_p(smc_drv_remove),
 	.suspend	= smc_drv_suspend,
 	.resume		= smc_drv_resume,
 	.driver		= {

@@ -2317,7 +2317,7 @@ static int __init smc_init(void)
 #endif
 #endif
 
-	return platform_driver_register(&smc_driver);
+	return platform_driver_probe(&smc_driver, smc_drv_probe);
 }
 
 static void __exit smc_cleanup(void)