Re: [PATCH] memstick: Fix memory leak in memstick_check() error path

[Catalin Marinas]

On 3 October 2013 22:13, Larry Finger <Larry.Finger@xxxxxxxxxxxx> wrote:
> With kernel 3.12-rc3, kemcheck reports the following leak:

That would be "kmemleak" rather than "kmemcheck" ;)

>> unreferenced object 0xffff8800ae85c190 (size 16):
>>   comm "kworker/u4:3", pid 685, jiffies 4294916336 (age 2831.760s)
>>   hex dump (first 16 bytes):
>>     6d 65 6d 73 74 69 63 6b 30 00 00 00 00 00 00 00  memstick0.......
>>   backtrace:
>>     [<ffffffff8146a0d1>] kmemleak_alloc+0x21/0x50
>>     [<ffffffff81160720>] __kmalloc_track_caller+0x160/0x2f0
>>     [<ffffffff81237b9b>] kvasprintf+0x5b/0x90
>>     [<ffffffff8122c0c1>] kobject_set_name_vargs+0x21/0x60
>>     [<ffffffff812e7f5c>] dev_set_name+0x3c/0x40
>>     [<ffffffffa02bf918>] memstick_check+0xb8/0x340 [memstick]
>>     [<ffffffff81069862>] process_one_work+0x1d2/0x670
>>     [<ffffffff8106a88a>] worker_thread+0x11a/0x370
>>     [<ffffffff81072ea6>] kthread+0xd6/0xe0
>>     [<ffffffff81478bbc>] ret_from_fork+0x7c/0xb0
>
> This problem was introduced by commit 0252c3b "memstick: struct device -
> replace bus_id with dev_name(), dev_set_name() where the name is not freed
> in the error path.
>
> Thanks to Catalin Marinas for suggesting the fix.
>
> Cc: Kay Sievers <kay.sievers@xxxxxxxx>
> Cc: Alex Dubov <oakad@xxxxxxxxx>
> Cc: Greg Kroah-Hartman <gregkh@xxxxxxx>
> Signed-off-by: Larry Finger <Larry.Finger@xxxxxxxxxxxx>
> ---
>  drivers/memstick/core/memstick.c | 1 +
>  1 file changed, 1 insertion(+)
>
> diff --git a/drivers/memstick/core/memstick.c b/drivers/memstick/core/memstick.c
> index ffcb10a..0c73a45 100644
> --- a/drivers/memstick/core/memstick.c
> +++ b/drivers/memstick/core/memstick.c
> @@ -415,6 +415,7 @@ static struct memstick_dev *memstick_alloc_card(struct memstick_host *host)
>         return card;
>  err_out:
>         host->card = old_card;
> +       kfree(card->dev.kobj.name);

It looks weird to go into dev.kobj internals here for freeing the
name. There is also memstick_free_card() which doesn't seem to do
anything about the name freeing.

Should memstick_alloc_card() do a device_initialise(&card->dev) and in
memstick_free_card() (or the error path) do a put_device(&card->dev)?
This should take care of kobj.name as well via kobject_put().

Catalin
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/