Patchwork rtc-cmos: do dev_set_drvdata() earlier in the initialization

login
register
mail settings
Submitter Dan Carpenter
Date May 17, 2010, 10:02 p.m.
Message ID <20100517220123.GA7691@bicker>
Download mbox | patch
Permalink /patch/52818/
State New
Headers show

Comments

Dan Carpenter - May 17, 2010, 10:02 p.m.
This is in response to https://bugzilla.kernel.org/show_bug.cgi?id=15963

The bug is an oops when dev_get_drvdata() returned null in
cmos_update_irq_enable().  The call tree looks like this:
  rtc_dev_ioctl()
    => rtc_update_irq_enable()
      => cmos_update_irq_enable()

It's caused by a race condition in the module initialization.  It is 
rtc_device_register() which makes the ioctl operations live so I moved
the call to dev_set_drvdata() before the call to rtc_device_register().

Reported-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Dan Carpenter <error27@gmail.com>
Tested-by: Randy Dunlap <randy.dunlap@oracle.com>

Patch

diff --git a/drivers/rtc/rtc-cmos.c b/drivers/rtc/rtc-cmos.c
index e9aa814..aa13875 100644
--- a/drivers/rtc/rtc-cmos.c
+++ b/drivers/rtc/rtc-cmos.c
@@ -719,6 +719,9 @@  cmos_do_probe(struct device *dev, struct resource *ports, int rtc_irq)
 		}
 	}
 
+	cmos_rtc.dev = dev;
+	dev_set_drvdata(dev, &cmos_rtc);
+
 	cmos_rtc.rtc = rtc_device_register(driver_name, dev,
 				&cmos_rtc_ops, THIS_MODULE);
 	if (IS_ERR(cmos_rtc.rtc)) {
@@ -726,8 +729,6 @@  cmos_do_probe(struct device *dev, struct resource *ports, int rtc_irq)
 		goto cleanup0;
 	}
 
-	cmos_rtc.dev = dev;
-	dev_set_drvdata(dev, &cmos_rtc);
 	rename_region(ports, dev_name(&cmos_rtc.rtc->dev));
 
 	spin_lock_irq(&rtc_lock);