Re: [PATCH 6/13: eCryptfs] Superblock operations

["Michael Thompson" <michael.craig.thompson@gmail.com>]

On 5/4/06, Pekka Enberg <penberg@cs.helsinki.fi> wrote:
> Hi Phillip,
>
> Some comments below.
>
> On 5/4/06, Phillip Hellewell <phillip@hellewell.homeip.net> wrote:
> > +kmem_cache_t *ecryptfs_inode_info_cache;
>
> Please use struct kmem_cache instead of the typedef.

Please explain why. Looking at the source shows that kmem_cache_t is
more widely used, and therefore seems to be the prefered way.

> > +       ecryptfs_printk(KERN_DEBUG, "Exit\n");
> > +}
> > +
> > +/**
> > + * Set up the ecryptfs inode.
> > + */
> > +static void ecryptfs_read_inode(struct inode *inode)
> > +{
> > +       ecryptfs_printk(KERN_DEBUG, "Enter; inode = [%p]\n", inode);
> > +       /* This is where we setup the self-reference in the vfs_inode's
> > +        * u.generic_ip. That way we don't have to walk the list again. */
> > +       ECRYPTFS_INODE_TO_PRIVATE_SM(inode) =
> > +               list_entry(inode, struct ecryptfs_inode_info, vfs_inode);
> > +       ECRYPTFS_INODE_TO_LOWER(inode) = NULL;
>
> Hmm, ugly, please make the setters explicit instead.

Curious, what exactly do you mean by this? I'm not sure what you mean
by "setters".

--
Michael C. Thompson <mcthomps@us.ibm.com>
Software-Engineer, IBM LTC Security