[Kernel-janitors] Re: [PATCH 2.6.9-rc2 1/4] w1/dscore: replace

[Evgeniy Polyakov <johnpol@2ka.mipt.ru>]

[-- Attachment #1.1: Type: text/plain, Size: 1450 bytes --]

On Tue, 2004-09-28 at 07:19, Evgeniy Polyakov wrote:
> On Tue, 2004-09-28 at 03:17, Nishanth Aravamudan wrote:
> > Any comments would be appreciated.
> 
> Following code is buggy, I know, the latest -mm tree already contains
> codepath which properly sets process state and checks signals.
> I've not used msleep*() due to future ability of special signal
> handling.

Hmm, I've rechecked -mm3, no it is not containing such fixes yet. I hope
next -mm release will have them.
Anyway if you set TASK_INTERRUPTIBLE you must be prepared to check
signal pending and either flush them or proper handle.

> > Description: Use msleep_interruptible() instead of schedule_timeout() to
> > guarantee the task delays as expected.
> > 
> > --- 2.6.9-rc2-vanilla/drivers/w1/dscore.c	2004-09-13 17:16:07.000000000 -0700
> > +++ 2.6.9-rc2/drivers/w1/dscore.c	2004-09-27 16:16:47.000000000 -0700
> > @@ -23,6 +23,7 @@
> >  #include <linux/kernel.h>
> >  #include <linux/mod_devicetable.h>
> >  #include <linux/usb.h>
> > +#include <linux/delay.h>
> >  
> >  #include "dscore.h"
> >  
> > @@ -731,7 +732,7 @@ void ds_disconnect(struct usb_interface 
> >  	usb_set_intfdata (intf, NULL);
> >  
> >  	while(atomic_read(&dev->refcnt))
> > -		schedule_timeout(HZ);
> > +		msleep_interruptible(1000);
> >  
> >  	usb_put_dev(dev->udev);
> >  	kfree(dev);
-- 
	Evgeniy Polyakov

Crash is better than data corruption. -- Art Grabowski

[-- Attachment #1.2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

[-- Attachment #2: Type: text/plain, Size: 167 bytes --]

_______________________________________________
Kernel-janitors mailing list
Kernel-janitors@lists.osdl.org
http://lists.osdl.org/mailman/listinfo/kernel-janitors