staging: Remove unnecessary *_set_drvdata()

Unnecessary [platform|pci|spi|dev]_set_drvdata() have been removed since the driver core clears the driver data to NULL after device release or on probe failure. There is no need to manually clear the device driver data to NULL.

The Coccinelle semantic patch used to make this change is as follows:
// <smpl>
@@
struct platform_device *pdev;
@@
- platform_set_drvdata(pdev, NULL);
@@
struct pci_dev *pci;
@@
- pci_set_drvdata(pci, NULL);
@@
struct spi_device *spi;
@@
- spi_set_drvdata(spi, NULL);
@@
struct device *dev;
@@
- dev_set_drvdata(dev, NULL);
@@
struct amba_device *dev;
@@
- amba_set_drvdata(dev, NULL);
// </smpl>

Amitoj Kaur Chawla (6):
staging: fbtft: Remove unnecessary spi_set_drvdata()
staging: sm750fb: Remove unnecessary pci_set_drvdata()
staging: media: omap4iss: Remove unnecessary platform_set_drvdata()
staging: unisys: visorbus: Remove unnecessary dev_set_drvdata()
staging: rdma: hfi1: Remove unnecessary pci_set_drvdata()
staging: rts5208: Remove unnecessary pci_set_drvdata()

 drivers/staging/fbtft/fbtft-core.c              | 5 -----
 drivers/staging/media/omap4iss/iss.c            | 2 --
 drivers/staging/rdma/hfi1/pcie.c                | 2 --
 drivers/staging/rts5208/rtsx.c                  | 2 --
 drivers/staging/sm750fb/sm750.c                 | 1 -
 drivers/staging/unisys/visorbus/visorbus_main.c | 1 -
 6 files changed, 13 deletions(-)

1.9.1

Unnecessary spi_set_drvdata() has been removed since the driver core clears the driver data to NULL after device release or on probe failure. There is no need to manually clear the device driver data to NULL.

The Coccinelle semantic patch used to make this change is as follows:
@@
struct spi_device *spi;
@@
- spi_set_drvdata(spi, NULL);
// </smpl>

Due to this removal, variable spi was unused so dropped spi

---
 drivers/staging/fbtft/fbtft-core.c | 5 -----
 1 file changed, 5 deletions(-)

index 6c9cdd2..574c148 100644
@@ -985,8 +985,6 @@ int fbtft_register_framebuffer(struct fb_info *fb_info)
 reg_fail:
 	if (par->fbtftops.unregister_backlight)
 		par->fbtftops.unregister_backlight(par);
-	if (spi)
-		spi_set_drvdata(spi, NULL);
 
 	return ret;
 }

@@ -1004,10 +1002,7 @@ EXPORT_SYMBOL(fbtft_register_framebuffer);
 int fbtft_unregister_framebuffer(struct fb_info *fb_info)
 {
 	struct fbtft_par *par = fb_info->par;
-	struct spi_device *spi = par->spi;
 
-	if (spi)
-		spi_set_drvdata(spi, NULL);
 	if (par->fbtftops.unregister_backlight)
 		par->fbtftops.unregister_backlight(par);
 	fbtft_sysfs_exit(par);

1.9.1

Unnecessary pci_set_drvdata() has been removed since the driver core clears the driver data to NULL after device release or on probe failure. There is no need to manually clear the device driver data to NULL.

The Coccinelle semantic patch used to make this change is as follows:
//<smpl>
@@
struct pci_dev *pci;
@@
- pci_set_drvdata(pci, NULL);
//</smpl>

---
 drivers/staging/sm750fb/sm750.c | 1 -
 1 file changed, 1 deletion(-)

index b889409..de996a6 100644
@@ -1135,7 +1135,6 @@ static void lynxfb_pci_remove(struct pci_dev *pdev)
 	iounmap(sm750_dev->pvMem);
 	kfree(g_settings);
 	kfree(sm750_dev);
-	pci_set_drvdata(pdev, NULL);
 }
 
 static int __init lynxfb_setup(char *options)

1.9.1

Unnecessary platform_set_drvdata() has been removed since the driver core clears the driver data to NULL after device release or on probe failure. There is no need to manually clear the device driver data to NULL.

The Coccinelle semantic patch used to make this change is as follows:
//<smpl>
@@
struct platform_device *pdev;
@@
- platform_set_drvdata(pdev, NULL);
//</smpl>

---
 drivers/staging/media/omap4iss/iss.c | 2 --
 1 file changed, 2 deletions(-)

index 30b473c..dea67d9 100644
@@ -1527,8 +1527,6 @@ error_modules:
 error_iss:
 	omap4iss_put(iss);
 error:
-	platform_set_drvdata(pdev, NULL);
-
 	mutex_destroy(&iss->iss_mutex);
 
 	return ret;

1.9.1

Unnecessary dev_set_drvdata() has been removed since the driver core clears the driver data to NULL after device release or on probe failure. There is no need to manually clear the device driver data to NULL.

The Coccinelle semantic patch used to make this change is as follows:
//<smpl>
@@
struct device *dev;
@@
- dev_set_drvdata(dev, NULL);
//</smpl>

---
 drivers/staging/unisys/visorbus/visorbus_main.c | 1 -
 1 file changed, 1 deletion(-)

index a301385..ec8ff85 100644
@@ -221,7 +221,6 @@ visorbus_release_busdevice(struct device *xdev)
 {
 	struct visor_device *dev = dev_get_drvdata(xdev);
 
-	dev_set_drvdata(xdev, NULL);
 	kfree(dev);
 }
 

1.9.1

Unnecessary pci_set_drvdata() has been removed since the driver core clears the driver data to NULL after device release or on probe failure. There is no need to manually clear the device driver data to NULL.

The Coccinelle semantic patch used to make this change is as follows:
//<smpl>
@@
struct pci_dev *pci;
@@
- pci_set_drvdata(pci, NULL);
//</smpl>

---
 drivers/staging/rdma/hfi1/pcie.c | 2 --
 1 file changed, 2 deletions(-)

index b2f553d8..47ca631 100644
@@ -247,8 +247,6 @@ void hfi1_pcie_ddcleanup(struct hfi1_devdata *dd)
 		iounmap(dd->rcvarray_wc);
 	if (dd->piobase)
 		iounmap(dd->piobase);
-
-	pci_set_drvdata(dd->pcidev, NULL);
 }
 
 /*

1.9.1

Unnecessary pci_set_drvdata() has been removed since the driver core clears the driver data to NULL after device release or on probe failure. There is no need to manually clear the device driver data to NULL.

The Coccinelle semantic patch used to make this change is as follows:
//<smpl>
@@
struct pci_dev *pci;
@@
- pci_set_drvdata(pci, NULL);
//</smpl>

---
 drivers/staging/rts5208/rtsx.c | 2 --
 1 file changed, 2 deletions(-)

index 7882193..b231946 100644
@@ -1001,8 +1001,6 @@ static void rtsx_remove(struct pci_dev *pci)
 
 	quiesce_and_remove_host(dev);
 	release_everything(dev);
-
-	pci_set_drvdata(pci, NULL);
 }
 
 /* PCI IDs */

1.9.1

In principle, your semantic patch should check whether the call is in the position of a device release or a probe failure. In practice though, it seems hard to imagine why this code would be in any place other than these cases.

julia

Okay, I'll modify the patch, other than that, should I resend the series with the modified patch?

Or just use the modified one for future cases?

Amitoj

Future is fine.

julia