Em Mon, 25 Feb 2013 01:39:41 +0200
Antti Palosaari <cr...@iki.fi> escreveu:

> WARNING: at lib/dma-debug.c:947 check_for_stack+0xa7/0xf0()
> ehci-pci 0000:00:04.1: DMA-API: device driver maps memory fromstack
> 
> Reported-by: poma <pomidorabelis...@gmail.com>
> Signed-off-by: Antti Palosaari <cr...@iki.fi>
> ---
>  drivers/media/usb/dvb-usb-v2/af9015.c | 34 ++++++++++++++++------------------
>  drivers/media/usb/dvb-usb-v2/af9015.h |  2 ++
>  2 files changed, 18 insertions(+), 18 deletions(-)
> 
> diff --git a/drivers/media/usb/dvb-usb-v2/af9015.c 
> b/drivers/media/usb/dvb-usb-v2/af9015.c
> index b86d0f2..28983aa 100644
> --- a/drivers/media/usb/dvb-usb-v2/af9015.c
> +++ b/drivers/media/usb/dvb-usb-v2/af9015.c
> @@ -30,22 +30,20 @@ DVB_DEFINE_MOD_OPT_ADAPTER_NR(adapter_nr);
>  
>  static int af9015_ctrl_msg(struct dvb_usb_device *d, struct req_t *req)
>  {
> -#define BUF_LEN 63
>  #define REQ_HDR_LEN 8 /* send header size */
>  #define ACK_HDR_LEN 2 /* rece header size */
>       struct af9015_state *state = d_to_priv(d);
>       int ret, wlen, rlen;
> -     u8 buf[BUF_LEN];
>       u8 write = 1;
>  
> -     buf[0] = req->cmd;
> -     buf[1] = state->seq++;
> -     buf[2] = req->i2c_addr;
> -     buf[3] = req->addr >> 8;
> -     buf[4] = req->addr & 0xff;
> -     buf[5] = req->mbox;
> -     buf[6] = req->addr_len;
> -     buf[7] = req->data_len;
> +     state->buf[0] = req->cmd;
> +     state->buf[1] = state->seq++;
> +     state->buf[2] = req->i2c_addr;
> +     state->buf[3] = req->addr >> 8;
> +     state->buf[4] = req->addr & 0xff;
> +     state->buf[5] = req->mbox;
> +     state->buf[6] = req->addr_len;
> +     state->buf[7] = req->data_len;
>  
>       switch (req->cmd) {
>       case GET_CONFIG:
> @@ -55,14 +53,14 @@ static int af9015_ctrl_msg(struct dvb_usb_device *d, 
> struct req_t *req)
>               break;
>       case READ_I2C:
>               write = 0;
> -             buf[2] |= 0x01; /* set I2C direction */
> +             state->buf[2] |= 0x01; /* set I2C direction */
>       case WRITE_I2C:
> -             buf[0] = READ_WRITE_I2C;
> +             state->buf[0] = READ_WRITE_I2C;
>               break;
>       case WRITE_MEMORY:
>               if (((req->addr & 0xff00) == 0xff00) ||
>                   ((req->addr & 0xff00) == 0xae00))
> -                     buf[0] = WRITE_VIRTUAL_MEMORY;
> +                     state->buf[0] = WRITE_VIRTUAL_MEMORY;
>       case WRITE_VIRTUAL_MEMORY:
>       case COPY_FIRMWARE:
>       case DOWNLOAD_FIRMWARE:
> @@ -90,7 +88,7 @@ static int af9015_ctrl_msg(struct dvb_usb_device *d, struct 
> req_t *req)
>       rlen = ACK_HDR_LEN;
>       if (write) {
>               wlen += req->data_len;
> -             memcpy(&buf[REQ_HDR_LEN], req->data, req->data_len);
> +             memcpy(&state->buf[REQ_HDR_LEN], req->data, req->data_len);
>       } else {
>               rlen += req->data_len;
>       }
> @@ -99,21 +97,21 @@ static int af9015_ctrl_msg(struct dvb_usb_device *d, 
> struct req_t *req)
>       if (req->cmd == DOWNLOAD_FIRMWARE || req->cmd == RECONNECT_USB)
>               rlen = 0;
>  
> -     ret = dvb_usbv2_generic_rw(d, buf, wlen, buf, rlen);
> +     ret = dvb_usbv2_generic_rw(d, state->buf, wlen, state->buf, rlen);
>       if (ret)
>               goto error;
>  
>       /* check status */
> -     if (rlen && buf[1]) {
> +     if (rlen && state->buf[1]) {
>               dev_err(&d->udev->dev, "%s: command failed=%d\n",
> -                             KBUILD_MODNAME, buf[1]);
> +                             KBUILD_MODNAME, state->buf[1]);
>               ret = -EIO;
>               goto error;
>       }
>  
>       /* read request, copy returned data to return buf */
>       if (!write)
> -             memcpy(req->data, &buf[ACK_HDR_LEN], req->data_len);
> +             memcpy(req->data, &state->buf[ACK_HDR_LEN], req->data_len);

Now that you're using just one buffer here, you'll need to protect this
routine with a mutex, as af9015_rc_query() may be called by a kthread 
while af9015_ctrl_msg() is running. As the RC polling code will also call
af9015_ctrl_msg() and both will be filling the very same state->buf, a
race condition happens.

-- 

Cheers,
Mauro
--
To unsubscribe from this list: send the line "unsubscribe linux-media" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Reply via email to