Re: [PATCH] rtlwifi: use %*ph[C] to dump small buffers

[Larry Finger <Larry.Finger@xxxxxxxxxxxx>]

On 09/26/2012 11:45 AM, Joe Perches wrote:
> On Wed, 2012-09-26 at 16:57 +0300, Andy Shevchenko wrote:
> > Signed-off-by: Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>
>
> Hi Andy.
>
> (adding Larry and Chaoming to cc's)
>
> Please use scripts/get_maintainer.pl on your patches to
> see who maintains these files so you can cc them.
>
> One comment below, it looks like a possible endian bug.
> (not in your patch)
>
> > ---
> >   drivers/net/wireless/rtlwifi/cam.c          |    7 ++-----
> >   drivers/net/wireless/rtlwifi/rtl8192ce/hw.c |    6 ++----
> >   drivers/net/wireless/rtlwifi/rtl8192cu/hw.c |    6 ++----
> >   3 files changed, 6 insertions(+), 13 deletions(-)
> >
> > diff --git a/drivers/net/wireless/rtlwifi/cam.c b/drivers/net/wireless/rtlwifi/cam.c
> > index 5b4b4d4..6ba9f80 100644
> > --- a/drivers/net/wireless/rtlwifi/cam.c
> > +++ b/drivers/net/wireless/rtlwifi/cam.c
> > @@ -52,11 +52,8 @@ static void rtl_cam_program_entry(struct ieee80211_hw *hw, u32 entry_no,
> >   	u32 target_content = 0;
> >   	u8 entry_i;
> >
> > -	RT_TRACE(rtlpriv, COMP_SEC, DBG_LOUD,
> > -		 "key_cont_128:\n %x:%x:%x:%x:%x:%x\n",
> > -		 key_cont_128[0], key_cont_128[1],
> > -		 key_cont_128[2], key_cont_128[3],
> > -		 key_cont_128[4], key_cont_128[5]);
> > +	RT_TRACE(rtlpriv, COMP_SEC, DBG_LOUD, "key_cont_128: %*ph\n",
> > +			6, key_cont_128);
> >
> >   	for (entry_i = 0; entry_i < CAM_CONTENT_COUNT; entry_i++) {
> >   		target_command = entry_i + CAM_CONTENT_COUNT * entry_no;
> > diff --git a/drivers/net/wireless/rtlwifi/rtl8192ce/hw.c b/drivers/net/wireless/rtlwifi/rtl8192ce/hw.c
> > index 86d73b3..932780d 100644
> > --- a/drivers/net/wireless/rtlwifi/rtl8192ce/hw.c
> > +++ b/drivers/net/wireless/rtlwifi/rtl8192ce/hw.c
> > @@ -1918,10 +1918,8 @@ static void rtl92ce_update_hal_rate_mask(struct ieee80211_hw *hw,
> >   				     (ratr_index << 28);
> >   	rate_mask[4] = macid | (shortgi ? 0x20 : 0x00) | 0x80;
> >   	RT_TRACE(rtlpriv, COMP_RATR, DBG_DMESG,
> > -		 "Rate_index:%x, ratr_val:%x, %x:%x:%x:%x:%x\n",
> > -		 ratr_index, ratr_bitmap,
> > -		 rate_mask[0], rate_mask[1], rate_mask[2], rate_mask[3],
> > -		 rate_mask[4]);
> > +		 "Rate_index:%x, ratr_val:%x, %*phC\n",
> > +		 ratr_index, ratr_bitmap, 5, rate_mask);
> >   	rtl92c_fill_h2c_cmd(hw, H2C_RA_MASK, 5, rate_mask);
> >
> >   	if (macid != 0)
> > diff --git a/drivers/net/wireless/rtlwifi/rtl8192cu/hw.c b/drivers/net/wireless/rtlwifi/rtl8192cu/hw.c
> > index 4bbb711..7554501 100644
> > --- a/drivers/net/wireless/rtlwifi/rtl8192cu/hw.c
> > +++ b/drivers/net/wireless/rtlwifi/rtl8192cu/hw.c
> > @@ -2169,10 +2169,8 @@ void rtl92cu_update_hal_rate_mask(struct ieee80211_hw *hw, u8 rssi_level)
> >   				      ratr_index << 28);
> >   	rate_mask[4] = macid | (shortgi ? 0x20 : 0x00) | 0x80;
> >   	RT_TRACE(rtlpriv, COMP_RATR, DBG_DMESG,
> > -		 "Rate_index:%x, ratr_val:%x, %x:%x:%x:%x:%x\n",
> > -		 ratr_index, ratr_bitmap,
> > -		 rate_mask[0], rate_mask[1], rate_mask[2], rate_mask[3],
> > -		 rate_mask[4]);
> > +		 "Rate_index:%x, ratr_val:%x, %*phC\n",
> > +		 ratr_index, ratr_bitmap, 5, rate_mask);
> >   	rtl92c_fill_h2c_cmd(hw, H2C_RA_MASK, 5, rate_mask);
> >   }
>
> rate_mask uses:
>
> 	u32 ratr_bitmap = (u32) mac->basic_rates;
> ...
> 	u8 rate_mask[5];
> ...
> 	[sets ratr_bitmap as u32]
> ...
> 	*(u32 *)&rate_mask = ((ratr_bitmap & 0x0fffffff) |
> 				      ratr_index << 28);
> ...
> 	rtl92c_fill_h2c_cmd(hw, H2C_RA_MASK, 5, rate_mask);
>
> Looks like a possible endian misuse to me.

After a second look at the code, I don't think this is an endian issue; however, 
I am proposing the following changes to remove any doubt:

Index: wireless-testing-new/drivers/net/wireless/rtlwifi/rtl8192ce/hw.c
===================================================================
--- wireless-testing-new.orig/drivers/net/wireless/rtlwifi/rtl8192ce/hw.c
+++ wireless-testing-new/drivers/net/wireless/rtlwifi/rtl8192ce/hw.c
@@ -1964,8 +1965,9 @@ static void rtl92ce_update_hal_rate_mask

        RT_TRACE(rtlpriv, COMP_RATR, DBG_DMESG,
                 "ratr_bitmap :%x\n", ratr_bitmap);
-       *(u32 *)&rate_mask = (ratr_bitmap & 0x0fffffff) |
-                                    (ratr_index << 28);
+       for (i = 0; i < 3; i++)
+               rate_mask[i] = ratr_bitmap & (0xff << (i * 4));
+       rate_mask[3] = (ratr_bitmap & 0x0f000000) | (ratr_index << 28);
        rate_mask[4] = macid | (shortgi ? 0x20 : 0x00) | 0x80;
        RT_TRACE(rtlpriv, COMP_RATR, DBG_DMESG,
                 "Rate_index:%x, ratr_val:%x, %*phC\n",

The downstream routine uses this info as an array of u8, thus it is OK. A proper 
patch will be sent later. At least we avoid that ugly cast.

Larry


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