On Tue, Oct 27, 2009 at 4:28 PM, David Vrabel <david.vrabel@xxxxxxx> wrote:
> sdio_read() may use DMA so read the Type-A header into a kmalloc'd
> buffer instead of an on-stack buffer (which results in a DMA API
> warning).
>
> Signed-off-by: David Vrabel <david.vrabel@xxxxxxx>
> ---
> Apply this instead, it frees the header buffer.
>
> David
>
> drivers/bluetooth/btsdio.c | 35 ++++++++++++++++++++++-------------
> 1 files changed, 22 insertions(+), 13 deletions(-)
>
> diff --git a/drivers/bluetooth/btsdio.c b/drivers/bluetooth/btsdio.c
> index 7e29827..59d0d3b 100644
> --- a/drivers/bluetooth/btsdio.c
> +++ b/drivers/bluetooth/btsdio.c
> @@ -54,6 +54,7 @@ MODULE_DEVICE_TABLE(sdio, btsdio_table);
> struct btsdio_data {
> struct hci_dev *hdev;
> struct sdio_func *func;
> + u8 *hdr_buf;
>
> struct work_struct work;
>
> @@ -122,7 +123,7 @@ static void btsdio_work(struct work_struct *work)
>
> static int btsdio_rx_packet(struct btsdio_data *data)
> {
> - u8 hdr[4] __attribute__ ((aligned(4)));
> + u8 *hdr = data->hdr_buf;
> struct sk_buff *skb;
> int err, len;
>
> @@ -292,6 +293,7 @@ static void btsdio_destruct(struct hci_dev *hdev)
>
> BT_DBG("%s", hdev->name);
>
> + kfree(data->hdr_buf);
> kfree(data);
> }
>
> @@ -299,9 +301,9 @@ static int btsdio_probe(struct sdio_func *func,
> const struct sdio_device_id *id)
> {
> struct btsdio_data *data;
> - struct hci_dev *hdev;
> + struct hci_dev *hdev = NULL;
This kind of assignments prevent useful warning from compiler, maybe
it's better to stage
the error bailout.
> struct sdio_func_tuple *tuple = func->tuples;
> - int err;
> + int err = -ENOMEM;
You are assuming no other code between 2 allocations, it's seems to
me error prone.
>
Thanks
Tomas
--
To unsubscribe from this list: send the line "unsubscribe linux-bluetooth" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
