Re: [RFC][patch 4/7] v9fs: VFS superblock operations (2.0-rc6)

[Pekka Enberg <penberg@gmail.com>]

Hi,

On 5/24/05, ericvh@gmail.com <ericvh@gmail.com> wrote:
> Index: fs/9p/v9fs.c
> ===================================================================
> --- /dev/null  (tree:0bf32353105286a5624aeea862d35a4bbae09851)
> +++ 178666ee376655ef8ec19a2ffc0490241b428110/fs/9p/v9fs.c  (mode:100644)
> @@ -0,0 +1,573 @@
> +/*
> +  * Fcall Slab Accounting
> +  */
> +
> +struct v9fs_slab {
> +       struct list_head list;
> +
> +       int size;
> +       kmem_cache_t *slab;
> +};
> +
> +static LIST_HEAD(v9fs_slab_list);

[snip]

> +
> +/**
> + * find_slab - look up a slab by size
> + * @size: size of slab data
> + *
> + */
> +
> +static inline kmem_cache_t *find_slab(int size)

Hmm? Why do you need this? If you're missing functionality from the
slab allocator, please put that in mm/slab.c, not your filesystem!

> +void v9fs_session_close(struct v9fs_session_info *v9ses)
> +{

[snip]

> +       if (v9ses->name) {
> +               kfree(v9ses->name);
> +       }

kfree() handles NULL pointers just fine, so please drop the redundant
check (here and in various other places too).

                       Pekka