Re: [rtc-linux] Re: rtc_cmos oops in cmos_rtc_ioctl

[Alessandro Zummo <alessandro.zummo@towertech.it>]

On Tue, 22 Sep 2009 15:15:26 -0300
Herton Ronaldo Krzesinski <herton@mandriva.com.br> wrote:

> 
> Something like this?:
> 
> From a94365843ab40a1904c4bc244af4e551f2f2aca9 Mon Sep 17 00:00:00 2001
> From: Herton Ronaldo Krzesinski <herton@mandriva.com.br>
> Date: Tue, 22 Sep 2009 13:46:00 -0300
> Subject: [PATCH] rtc-cmos: convert RTC_AIE/RTC_UIE to rtc irq API
> 
> Drop ioctl function that handles RTC_AIE/RTC_UIE, and use instead the
> rtc subsystem API (alarm_irq_enable/update_irq_enable callbacks).
> 
> Signed-off-by: Herton Ronaldo Krzesinski <herton@mandriva.com.br>

 Looks nice. David?

> --
> Still, couldn't an oops still trigger with this with dev_set_drvdata after
> rtc_device_register? Lets say, a small test program like this:

 dev_set_drvdata should be called before rtc_device_register. I guess
 I'll have to patch every single driver :(


 [...]

> Compiled and run with similar udev rule. It's very unlikely, and I couldn't
> get an oops anymore unless I moved dev_set_drvdata down in cmos_do_probe in
> my test machine, but should be a possibility?

 I don't think so.
 


-- 

 Best regards,

 Alessandro Zummo,
  Tower Technologies - Torino, Italy

  http://www.towertech.it