On Wed, May 19, 2010 at 11:13:50AM -0700, H. Peter Anvin wrote:
> On 05/19/2010 11:08 AM, Frederic Weisbecker wrote:
> > On Wed, May 19, 2010 at 11:02:04AM -0700, H. Peter Anvin wrote:
> >> On 05/19/2010 10:24 AM, Frederic Weisbecker wrote:
> >>> * generate kernel reactions
> >>> */
> >>> -static int autofs_root_ioctl(struct inode *inode, struct file *filp,
> >>> +static int autofs_root_ioctl_unlocked(struct inode *inode, struct file
> >>> *filp,
> >>> unsigned int cmd, unsigned long arg)
> >>> {
> >>> struct autofs_sb_info *sbi = autofs_sbi(inode->i_sb);
> >>> @@ -579,3 +579,16 @@ static int autofs_root_ioctl(struct inode *inode,
> >>> struct file *filp,
> >>> return -ENOSYS;
> >>> }
> >>> }
> >>> +
> >>> +static long autofs_root_ioctl(struct file *filp,
> >>> + unsigned int cmd, unsigned long arg)
> >>> +{
> >>
> >> The choice of naming here seems reverse in my opinion.
> >
> >
> > Oh, why?
> >
> > The function that holds the bkl calls its unlocked version.
> >
>
> But it's not ... it is locked at that point. It's not lock*ing*, but it
> is not *unlocked*, either. Furthermore, it is directly contradicting
> the naming scheme of the ops structure.
>
> -hpa
>
Would you prefer me to resend a patch?
_______________________________________________
autofs mailing list
[email protected]
http://linux.kernel.org/mailman/listinfo/autofs
