On 23/06/12 10:46, Dan Carpenter wrote:
On Wed, Jun 20, 2012 at 01:36:10PM +0100, Rupesh Gujare wrote:
Register ozmo_wpan class with sysfs & support for udev
to create device node.
Signed-off-by: Rupesh Gujare <rgujare@xxxxxxxxxxxxxxx>
---
drivers/staging/ozwpan/ozcdev.c | 24 +++++++++++++++++++++++-
1 files changed, 23 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/ozwpan/ozcdev.c b/drivers/staging/ozwpan/ozcdev.c
index 27325f7..929756a 100644
--- a/drivers/staging/ozwpan/ozcdev.c
+++ b/drivers/staging/ozwpan/ozcdev.c
@@ -330,10 +330,12 @@ const struct file_operations oz_fops = {
int oz_cdev_register(void)
{
int err;
+ struct class *cl;
+ struct device *dev;
Put a blank line between declarations and statements.
memset(&g_cdev, 0, sizeof(g_cdev));
err = alloc_chrdev_region(&g_cdev.devnum, 0, 1, "ozwpan");
if (err < 0)
- return err;
+ goto out3;
Just return directly here. Don't use GW-BASIC names for labels.
The label name should reflect the what happens at the label. For
example: goto unregister; would meant that you unregister after the
goto.
oz_trace("Alloc dev number %d:%d\n", MAJOR(g_cdev.devnum),
MINOR(g_cdev.devnum));
cdev_init(&g_cdev.cdev, &oz_fops);
@@ -342,7 +344,27 @@ int oz_cdev_register(void)
spin_lock_init(&g_cdev.lock);
init_waitqueue_head(&g_cdev.rdq);
err = cdev_add(&g_cdev.cdev, g_cdev.devnum, 1);
+ if (err < 0) {
+ oz_trace("Failed to add cdev\n");
+ goto out2;
+ }
+ cl = class_create(THIS_MODULE, "ozmo_wpan");
+ if (IS_ERR(cl)) {
Needs a:
err = PTR_ERR(cl);
+ oz_trace("Failed to register ozmo_wpan class\n");
+ goto out1;
+ }
+ dev = device_create(cl, NULL, g_cdev.devnum, NULL, "ozwpan");
+ if (IS_ERR(dev)) {
+ oz_trace("Failed to create sysfs entry for cdev\n");
Needs a:
err = PTR_ERR(dev);
+ goto out1;
+ }
return 0;
Blank line here probably.
regards,
dan carpenter
+out1:
+ cdev_del(&g_cdev.cdev);
+out2:
+ unregister_chrdev_region(g_cdev.devnum, 1);
+out3:
+ return err;
}
/*------------------------------------------------------------------------------
* Context: process
--
Thanks Dan,
Sounds reasonable.
I will include these changes in next patch series.
--
Regards,
Rupesh Gujare
_______________________________________________
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxx
http://driverdev.linuxdriverproject.org/mailman/listinfo/devel
