Re: [PATCH] [v2] nvdimm-btt: fix several memleaks

[Ira Weiny <ira.weiny@intel.com>]

Dinghao Liu wrote:
> Resources allocated by kcalloc() in btt_freelist_init(),
> btt_rtt_init(), and btt_maplocks_init() are not correctly
> released in their callers when an error happens. For
> example, when an error happens in btt_freelist_init(), its
> caller discover_arenas() will directly free arena, which makes
> arena->freelist a leaked memory. Fix these memleaks by using
> devm_kcalloc() to make the memory auto-freed on driver detach.

Thanks for this work!

> 
> Fixes: 5212e11fde4d ("nd_btt: atomic sector updates")
> Signed-off-by: Dinghao Liu <dinghao.liu@zju.edu.cn>
> ---
> 
> Changelog:
> 
> v2: -Use devm_kcalloc() to fix the memleaks.
>     -Fix the potential leaked memory in btt_rtt_init()
>      and btt_maplocks_init().
> ---
>  drivers/nvdimm/btt.c | 35 ++++++++++++++++-------------------
>  1 file changed, 16 insertions(+), 19 deletions(-)
> 
> diff --git a/drivers/nvdimm/btt.c b/drivers/nvdimm/btt.c
> index d5593b0dc700..c55231f42617 100644
> --- a/drivers/nvdimm/btt.c
> +++ b/drivers/nvdimm/btt.c
> @@ -531,13 +531,13 @@ static int arena_clear_freelist_error(struct arena_info *arena, u32 lane)
>  	return ret;
>  }
>  
> -static int btt_freelist_init(struct arena_info *arena)
> +static int btt_freelist_init(struct device *dev, struct arena_info *arena)

Both struct arena_info and struct btt contain references to struct nd_btt
which is the device you are passing down this call stack.

Just use the device in the arena/btt rather than passing a device
structure.  That makes the code easier to read and ensures that the device
associated with this arena or btt is used.

>  {
>  	int new, ret;
>  	struct log_entry log_new;
>  	u32 i, map_entry, log_oldmap, log_newmap;
>  
> -	arena->freelist = kcalloc(arena->nfree, sizeof(struct free_entry),
> +	arena->freelist = devm_kcalloc(dev, arena->nfree, sizeof(struct free_entry),

	... devm_kcalloc(&arena->nd_btt.dev, ...)

>  					GFP_KERNEL);
>  	if (!arena->freelist)
>  		return -ENOMEM;
> @@ -718,20 +718,20 @@ static int log_set_indices(struct arena_info *arena)
>  	return 0;
>  }
>  
> -static int btt_rtt_init(struct arena_info *arena)
> +static int btt_rtt_init(struct device *dev, struct arena_info *arena)
>  {
> -	arena->rtt = kcalloc(arena->nfree, sizeof(u32), GFP_KERNEL);
> +	arena->rtt = devm_kcalloc(dev, arena->nfree, sizeof(u32), GFP_KERNEL);
>  	if (arena->rtt == NULL)
>  		return -ENOMEM;
>  
>  	return 0;
>  }
>  
> -static int btt_maplocks_init(struct arena_info *arena)
> +static int btt_maplocks_init(struct device *dev, struct arena_info *arena)
>  {
>  	u32 i;
>  
> -	arena->map_locks = kcalloc(arena->nfree, sizeof(struct aligned_lock),
> +	arena->map_locks = devm_kcalloc(dev, arena->nfree, sizeof(struct aligned_lock),
>  				GFP_KERNEL);
>  	if (!arena->map_locks)
>  		return -ENOMEM;
> @@ -805,9 +805,6 @@ static void free_arenas(struct btt *btt)
>  
>  	list_for_each_entry_safe(arena, next, &btt->arena_list, list) {
>  		list_del(&arena->list);
> -		kfree(arena->rtt);
> -		kfree(arena->map_locks);
> -		kfree(arena->freelist);

This does not quite work.

free_arenas() is used in the error paths of create_arenas() and
discover_arenas().  In those cases devm_kfree() is probably a better way
to clean up this.

However...

>  		debugfs_remove_recursive(arena->debugfs_dir);
>  		kfree(arena);

Why can't arena be allocated with devm_*?

We need to change this up a bit more to handle the error path vs regular
device shutdown free (automatic devm frees).

Ira