linux-kernel - Re: [RFC] staging: vt6656: Add formula to the vnt_rf

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Wed, 15 Apr 2020 18:25:41 +0200
From:   Oscar Carter <oscar.carter@....com>
To:     Dan Carpenter <dan.carpenter@...cle.com>
Cc:     Oscar Carter <oscar.carter@....com>,
        Forest Bond <forest@...ttletooquiet.net>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Malcolm Priestley <tvboxspy@...il.com>,
        devel@...verdev.osuosl.org, linux-kernel@...r.kernel.org
Subject: Re: [RFC] staging: vt6656: Add formula to the vnt_rf_addpower
 function

On Tue, Apr 14, 2020 at 04:12:14PM +0300, Dan Carpenter wrote:
> On Mon, Apr 13, 2020 at 04:02:09PM +0200, Oscar Carter wrote:
> > diff --git a/drivers/staging/vt6656/rf.c b/drivers/staging/vt6656/rf.c
> > index 4f9aba0f21b0..3b200d7290a5 100644
> > --- a/drivers/staging/vt6656/rf.c
> > +++ b/drivers/staging/vt6656/rf.c
> > @@ -575,28 +575,14 @@ int vnt_rf_setpower(struct vnt_private *priv, u32 rate, u32 channel)
> >
> >  static u8 vnt_rf_addpower(struct vnt_private *priv)
> >  {
> > +	s32 base;
>
> Just use "int".  s32 is for when signed 32 bit is specified in the
> hardware.  I realize that it's done in this file, but if all your
> friends jumped off a bridge doesn't mean you should drink their kool-aid.

Ok, lesson learned and thanks for the aclaration about when use every type.

> >  	s32 rssi = -priv->current_rssi;
> >
> >  	if (!rssi)
> >  		return 7;
> >
> > -	if (priv->rf_type == RF_VT3226D0) {
> > -		if (rssi < -70)
> > -			return 9;
> > -		else if (rssi < -65)
> > -			return 7;
> > -		else if (rssi < -60)
> > -			return 5;
> > -	} else {
> > -		if (rssi < -80)
> > -			return 9;
> > -		else if (rssi < -75)
> > -			return 7;
> > -		else if (rssi < -70)
> > -			return 5;
> > -	}
> > -
> > -	return 0;
> > +	base = (priv->rf_type == RF_VT3226D0) ? -60 : -70;
> > +	return (rssi < base--) ? ((rssi - base) / -5) * 2 + 5 : 0;
>                        ^^^^^^
> I quite hate this postop.  It would have been cleaner to write it like:
>
> 	return (rssi < base) ? ((rssi - (base - 1)) / -5) * 2 + 5 : 0
                                        ^        ^
Now, if we apply the minus operator one parentheses can be removed. The
same expression is now:

  	return (rssi < base) ? ((rssi - base + 1) / -5) * 2 + 5 : 0

I think it's clear enought.

> I'm sorry, I'm not clever enough to figure out the potential values of
> "rssi".

The IEEE 802.11 standard specifies that RSSI can be on a scale of 0 to
up to 255, and that each chipset manufacturer can define their own max
RSSI value.  It's all up to the manufacturer.

> How did you work out this formula?  It feels like it came from
> a standard or something?

I realized that the two branches of the if statement return the same
values (5, 7, 9) and that each value has a difference of 2 units from
the previous one. Also, every branch has 3 ranges, and every range has
an interval of 5. The only difference in this case is the "base" value
of each branch.

So, the solution was obtain the range index --> (rssi - base) / -5
Then, we need two units for every range index -> * 2
Now, the return value starts with five -------> + 5

The base-- was to obtain the range index the same that the orignal
function.

> Do we not have a function already which implements the standard?

I have been searching but I have not found anything that relates the
RSSI value with the amount of power to add. I have found

struct station_parameters -> member txpwr (struct sta_txpwr type)

but all the functions related to this doesn't set the tx power
depending on the RSSI value.

> regards,
> dan carpenter
>

thanks,
oscar carter