Re: [PATCH 1/2] staging: wilc1000: fix some endianness sparse warnings

[Thibaut Robert]

Le mercredi 30 mai 2018 à 14:17:25 (+0300), Dan Carpenter a écrit :
> On Tue, May 29, 2018 at 09:11:43PM +0200, Thibaut Robert wrote:
> > diff --git a/drivers/staging/wilc1000/wilc_wfi_cfgoperations.c b/drivers/staging/wilc1000/wilc_wfi_cfgoperations.c
> > index e248702ee519..745bf5ca2622 100644
> > --- a/drivers/staging/wilc1000/wilc_wfi_cfgoperations.c
> > +++ b/drivers/staging/wilc1000/wilc_wfi_cfgoperations.c
> > @@ -1431,7 +1431,7 @@ void wilc_wfi_p2p_rx(struct net_device *dev, u8 *buff, u32 size)
> >  
> >  	freq = ieee80211_channel_to_frequency(curr_channel, NL80211_BAND_2GHZ);
> >  
> > -	if (!ieee80211_is_action(buff[FRAME_TYPE_ID])) {
> > +	if (!ieee80211_is_action(cpu_to_le16(buff[FRAME_TYPE_ID]))) {
> 
> "buff" comes from the network, it's going to be little endian, not cpu
> endian.  The rest of the function treats it as CPU endian but I'm pretty
> sure it's wrong...
buff comes from the network but we are looking at single byte here.
ieee80211_is_action expects an le16, so we I added this to extend an u8
to an le16. Is this incorrect ?

Or maybe we the buff has the second byte ? but that I can't tell. 

> 
> >  		cfg80211_rx_mgmt(priv->wdev, freq, 0, buff, size, 0);
> >  		return;
> >  	}
> 
> 
> > diff --git a/drivers/staging/wilc1000/wilc_wlan.c b/drivers/staging/wilc1000/wilc_wlan.c
> > index 28c93f3f846e..a5ac1d26590b 100644
> > --- a/drivers/staging/wilc1000/wilc_wlan.c
> > +++ b/drivers/staging/wilc1000/wilc_wlan.c
> > @@ -560,7 +560,8 @@ int wilc_wlan_handle_txq(struct net_device *dev, u32 *txq_count)
> >  	int ret = 0;
> >  	int counter;
> >  	int timeout;
> > -	u32 vmm_table[WILC_VMM_TBL_SIZE];
> > +	__le32 vmm_table[WILC_VMM_TBL_SIZE];
> > +	u32 table_entry;
> >  	struct wilc_vif *vif;
> >  	struct wilc *wilc;
> >  	const struct wilc_hif_func *func;
> > @@ -598,10 +599,10 @@ int wilc_wlan_handle_txq(struct net_device *dev, u32 *txq_count)
> >  			if ((sum + vmm_sz) > LINUX_TX_SIZE)
> >  				break;
> >  
> > -			vmm_table[i] = vmm_sz / 4;
> > +			table_entry = vmm_sz / 4;
> >  			if (tqe->type == WILC_CFG_PKT)
> > -				vmm_table[i] |= BIT(10);
> > -			vmm_table[i] = cpu_to_le32(vmm_table[i]);
> > +				table_entry |= BIT(10);
> > +			vmm_table[i] = cpu_to_le32(table_entry);
> >  
> >  			i++;
> >  			sum += vmm_sz;
> > @@ -704,8 +705,7 @@ int wilc_wlan_handle_txq(struct net_device *dev, u32 *txq_count)
> >  		if (vmm_table[i] == 0)
> >  			break;
> >  
> > -		vmm_table[i] = cpu_to_le32(vmm_table[i]);
> > -		vmm_sz = (vmm_table[i] & 0x3ff);
> > +		vmm_sz = (le32_to_cpu(vmm_table[i]) & 0x3ff);
> >  		vmm_sz *= 4;
> >  		header = (tqe->type << 31) |
> >  			 (tqe->buffer_size << 15) |
> > @@ -715,8 +715,7 @@ int wilc_wlan_handle_txq(struct net_device *dev, u32 *txq_count)
> >  		else
> >  			header &= ~BIT(30);
> >  
> > -		header = cpu_to_le32(header);
> > -		memcpy(&txb[offset], &header, 4);
> > +		*((__le32 *)&txb[offset]) = cpu_to_le32(header);
> 
> I worry about alignment issues here.  That might be the reason for the
> memcpy().  (I'm reading as fast as I can and don't the code so I may
> be wrong).
> 
> >  		if (tqe->type == WILC_CFG_PKT) {
> >  			buffer_offset = ETH_CONFIG_PKT_HDR_OFFSET;
> >  		} else if (tqe->type == WILC_NET_PKT) {
> > @@ -770,8 +769,7 @@ static void wilc_wlan_handle_rx_buff(struct wilc *wilc, u8 *buffer, int size)
> >  
> >  	do {
> >  		buff_ptr = buffer + offset;
> > -		memcpy(&header, buff_ptr, 4);
> > -		header = cpu_to_le32(header);
> > +		header = le32_to_cpup((__le32 *)buff_ptr);
> 
> Maybe the same, whenever you see a memcpy().
> 
> >  
> >  		is_cfg_packet = (header >> 31) & 0x1;
> >  		pkt_offset = (header >> 22) & 0x1ff;
> > @@ -942,6 +940,7 @@ int wilc_wlan_firmware_download(struct wilc *wilc, const u8 *buffer,
> >  	u32 offset;
> >  	u32 addr, size, size2, blksz;
> >  	u8 *dma_buffer;
> > +	const __le32 *header;
> >  	int ret = 0;
> >  
> >  	blksz = BIT(12);
> > @@ -952,10 +951,9 @@ int wilc_wlan_firmware_download(struct wilc *wilc, const u8 *buffer,
> >  
> >  	offset = 0;
> >  	do {
> > -		memcpy(&addr, &buffer[offset], 4);
> > -		memcpy(&size, &buffer[offset + 4], 4);
> > -		addr = cpu_to_le32(addr);
> > -		size = cpu_to_le32(size);
> > +		header = (__le32 *)buffer + offset;
> > +		addr = le32_to_cpu(header[0]);
> > +		size = le32_to_cpu(header[1]);
> >  		acquire_bus(wilc, ACQUIRE_ONLY);
> >  		offset += 8;
> >  		while (((int)size) && (offset < buffer_size)) {
> > diff --git a/drivers/staging/wilc1000/wilc_wlan_cfg.c b/drivers/staging/wilc1000/wilc_wlan_cfg.c
> > index c0b9b700f4d7..4a914d8572aa 100644
> > --- a/drivers/staging/wilc1000/wilc_wlan_cfg.c
> > +++ b/drivers/staging/wilc1000/wilc_wlan_cfg.c
> > @@ -275,14 +275,14 @@ static int wilc_wlan_cfg_set_bin(u8 *frame, u32 offset, u16 id, u8 *b, u32 size)
> >  
> >  static void wilc_wlan_parse_response_frame(u8 *info, int size)
> >  {
> > -	u32 wid, len = 0, i = 0;
> > +	u32 wid;
> > +	int len = 0, i = 0;
> 
> Why did we make these int now?
> 
> >  
> >  	while (size > 0) {
> >  		i = 0;
> > -		wid = info[0] | (info[1] << 8);
> > -		wid = cpu_to_le32(wid);
> > +		wid = le16_to_cpup((__le16 *)info);
> >  
> > -		switch ((wid >> 12) & 0x7) {
> > +		switch (info[1] >> 4) {
> 
> Why do we not need to mask by 0x7?  Anyway, I feel like this isn't
> beautiful.  We should be using a macro and "wid" instead of magically
> poking into info[1].
> 
> 		switch(SOME_MACRO(wid)) {
> 
> 
> regards,
> dan carpenter