[RFC] staging: r8188eu: Fix sparse warnings in core/rtw_br_ext.c

[RFC] staging: r8188eu: Fix sparse warnings in core/rtw_br_ext.c

[Larry Finger]

Please tell me if the following fixes are appropriate. In particular, I am uncertain of
the __force attributes.

Thanks,

Larry


Sparse lists the following:

  CHECK   drivers/staging/rtl8188eu/core/rtw_br_ext.c
drivers/staging/rtl8188eu/core/rtw_br_ext.c:1156:63: warning: restricted __sum16 degrades to integer
drivers/staging/rtl8188eu/core/rtw_br_ext.c:1160:69: warning: incorrect type in assignment (different base types)
drivers/staging/rtl8188eu/core/rtw_br_ext.c:1160:69:    expected restricted __sum16 [usertype] check
drivers/staging/rtl8188eu/core/rtw_br_ext.c:1160:69:    got int

---
 drivers/staging/rtl8188eu/core/rtw_br_ext.c | 47 +++++++++++++++--------------
 1 file changed, 24 insertions(+), 23 deletions(-)

diff --git a/drivers/staging/rtl8188eu/core/rtw_br_ext.c b/drivers/staging/rtl8188eu/core/rtw_br_ext.c
index 5425bc5..54a706f 100644
--- a/drivers/staging/rtl8188eu/core/rtw_br_ext.c
+++ b/drivers/staging/rtl8188eu/core/rtw_br_ext.c

@@ -1153,11 +1154,11 @@ void dhcp_flag_bcast(struct adapter *priv, struct sk_buff *skb)
 							/*  or BROADCAST flag */
 							dhcph->flags |= htons(BROADCAST_FLAG);
 							/*  recalculate checksum */
-							sum = ~(udph->check) & 0xffff;
+							sum = (__force int)(~(udph->check)) & 0xffff;
 							sum += be16_to_cpu(dhcph->flags);
 							while (sum >> 16)
 								sum = (sum & 0xffff) + (sum >> 16);
-							udph->check = ~sum;
+							udph->check = (__force __sum16)(~sum);
 						}
 					}
 				}
-- 
1.8.4

Re: [RFC] staging: r8188eu: Fix sparse warnings in core/rtw_br_ext.c

[Greg KH]

On Sat, Dec 21, 2013 at 07:31:08PM -0600, Larry Finger wrote:
> Please tell me if the following fixes are appropriate. In particular, I am uncertain of
> the __force attributes.
> 
> Thanks,
> 
> Larry
> 
> 
> Sparse lists the following:
> 
>   CHECK   drivers/staging/rtl8188eu/core/rtw_br_ext.c
> drivers/staging/rtl8188eu/core/rtw_br_ext.c:1156:63: warning: restricted __sum16 degrades to integer
> drivers/staging/rtl8188eu/core/rtw_br_ext.c:1160:69: warning: incorrect type in assignment (different base types)
> drivers/staging/rtl8188eu/core/rtw_br_ext.c:1160:69:    expected restricted __sum16 [usertype] check
> drivers/staging/rtl8188eu/core/rtw_br_ext.c:1160:69:    got int
> 
> ---
>  drivers/staging/rtl8188eu/core/rtw_br_ext.c | 47 +++++++++++++++--------------
>  1 file changed, 24 insertions(+), 23 deletions(-)
> 
> diff --git a/drivers/staging/rtl8188eu/core/rtw_br_ext.c b/drivers/staging/rtl8188eu/core/rtw_br_ext.c
> index 5425bc5..54a706f 100644
> --- a/drivers/staging/rtl8188eu/core/rtw_br_ext.c
> +++ b/drivers/staging/rtl8188eu/core/rtw_br_ext.c
> 
> @@ -1153,11 +1154,11 @@ void dhcp_flag_bcast(struct adapter *priv, struct sk_buff *skb)
>  							/*  or BROADCAST flag */
>  							dhcph->flags |= htons(BROADCAST_FLAG);
>  							/*  recalculate checksum */
> -							sum = ~(udph->check) & 0xffff;
> +							sum = (__force int)(~(udph->check)) & 0xffff;
>  							sum += be16_to_cpu(dhcph->flags);
>  							while (sum >> 16)
>  								sum = (sum & 0xffff) + (sum >> 16);
> -							udph->check = ~sum;
> +							udph->check = (__force __sum16)(~sum);

I thought __force was only "needed" when trying to move between user and
kernel pointer types, I didn't think it was needed for stuff like this.

But I really don't know what __sum16 is, so I can't say for sure, sorry.

greg k-h