Re: [PATCHv3 02/15] staging: rtl8192u: r8192U_core: add temporary variables to keep lines under 80 characters

From: RaphaÃl Beamonte
Date: Mon Sep 21 2015 - 12:31:39 EST


2015-09-20 22:16 GMT-04:00 Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>:
> On Sun, Sep 20, 2015 at 01:14:14PM -0400, RaphaÃl Beamonte wrote:
>> Add some temporary variables to reduce line length under the maximum
>> of 80 characters, as per the kernel code style.
>>
>> Signed-off-by: RaphaÃl Beamonte <raphael.beamonte@xxxxxxxxx>
>> ---
>> drivers/staging/rtl8192u/r8192U_core.c | 139 ++++++++++++++++++++++-----------
>> 1 file changed, 94 insertions(+), 45 deletions(-)
>>
>> diff --git a/drivers/staging/rtl8192u/r8192U_core.c b/drivers/staging/rtl8192u/r8192U_core.c
>> index 28b54ba..2abc3e77 100644
>> --- a/drivers/staging/rtl8192u/r8192U_core.c
>> +++ b/drivers/staging/rtl8192u/r8192U_core.c
>> @@ -171,6 +171,7 @@ static void rtl819x_set_channel_map(u8 channel_plan, struct r8192_priv *priv)
>> {
>> int i, max_chan = -1, min_chan = -1;
>> struct ieee80211_device *ieee = priv->ieee80211;
>> + struct CHANNEL_LIST cl;
>
> A whole structure on the stack? Are you sure you want to do that?

Nope. I don't want to. I don't know what I was thinking, or if even I
was. Sorry for that, should have been a pointer.

>>
>> switch (channel_plan) {
>> case COUNTRY_CODE_FCC:
>> @@ -194,15 +195,18 @@ static void rtl819x_set_channel_map(u8 channel_plan, struct r8192_priv *priv)
>> "unknown rf chip, can't set channel map in function:%s()\n",
>> __func__);
>> }
>> - if (ChannelPlan[channel_plan].Len != 0) {
>> + cl = ChannelPlan[channel_plan];
>
> You just did a memory copy of a whole structure, did you really want to
> do that? If so, why? You just changed the logic of this function,
> potentially slowing things down and setting yourself up for causing real
> bugs (hint, anything you now change in that structure is not going to be
> ever saved, it will go away after the function is exited.)

You're right. Should have been a pointer. Sorry again, that wasn't the
intention.

> Please be more aware of how structures and pointers work in C before
> doing changes like this, I can't take this change.
>
>> + if (cl.Len != 0) {
>> /* Clear old channel map */
>> memset(GET_DOT11D_INFO(ieee)->channel_map, 0,
>> sizeof(GET_DOT11D_INFO(ieee)->channel_map));
>> /* Set new channel map */
>> - for (i = 0; i < ChannelPlan[channel_plan].Len; i++) {
>> - if (ChannelPlan[channel_plan].Channel[i] < min_chan || ChannelPlan[channel_plan].Channel[i] > max_chan)
>> + for (i = 0; i < cl.Len; i++) {
>> + u8 chan = cl.Channel[i];
>> +
>> + if (chan < min_chan || chan > max_chan)
>> break;
>> - GET_DOT11D_INFO(ieee)->channel_map[ChannelPlan[channel_plan].Channel[i]] = 1;
>> + GET_DOT11D_INFO(ieee)->channel_map[chan] = 1;
>> }
>> }
>> break;
>> @@ -1649,9 +1653,12 @@ short rtl8192_tx(struct net_device *dev, struct sk_buff *skb)
>> &zero, 0, tx_zero_isr, dev);
>> status = usb_submit_urb(tx_urb_zero, GFP_ATOMIC);
>> if (status) {
>> + atomic_t tx =
>> + priv->tx_pending[tcb_desc->queue_index];
>
> Oh that's funny, think about what you just did here.

Also should have been a pointer. I'll do a more thorough recheck of
that patch before re-sending it. Thanks for your review and your time!

> Please get some more experience with C before doing kernel development.
> C is tricky and will let you shoot yourself in the foot and any other
> body part if you are not _VERY_ careful. You just blew off a few limbs
> without realizing it at all in just the first two changes you made.
>
> thanks,
>
> greg k-h
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/