Re: [PATCH] staging: unisys: handle major number properly

[Dan Carpenter <dan.carpenter@oracle.com>]

On Tue, Mar 24, 2015 at 02:13:16PM +0530, Sudip Mukherjee wrote:
> On Tue, Mar 24, 2015 at 11:32:47AM +0300, Dan Carpenter wrote:
> <snip>
> > >  	}
> > > -	rc = cdev_add(&file_cdev, MKDEV(MAJOR(majordev), 0), 1);
> > > +	rc = cdev_add(&file_cdev, MKDEV(major, 0), 1);
> > 
> > This should just be:
> > 
> > 	rc = cdev_add(&file_cdev, majordev, 1);
> > 
> > So here is my suggestion:
> > 
> > [patch 1] delete dead code I mentioned in my previous email.
> > 	This deletes "registered" and the (MAJOR(majordev) >= 0) check.
> 
> that was initially my first patch.

Yes, but the patch description wasn't clear.  You were talking saying
the condition was always true because of the types and I am saying the
condition is always true because of the order the functions are called.

> 
> > 
> > [patch 2] pass majordev to visorchipset_file_cleanup()
> > 	This lets you delete the "majordev" global.
> > 
> > [patch 3] small cleanup in visorchipset_file_init()
> > 
> > -	rc = cdev_add(&file_cdev, MKDEV(MAJOR(majordev), 0), 1);
> > +	rc = cdev_add(&file_cdev, majordev, 1);
> 
> and i can also include the removal of that global variable in this
> 3rd patch.
> thanks.. after Greg's review i was thinking i understood the code wrong.
> but then will this be a v2 or a whole new series?

I'm pretty sure we're up to v3 at least now.  This means redoing the
whole series, yes.

regards,
dan carpenter