[altq 1067] Re: Accuracy of atobps(const char*)

["Sankaran Narayanan" <sankaran@cs.columbia.edu>]

well, Kbps is kilo bits per second...the general convention is 1000 when
bits are involved, and 1024 (2^10) when bytes are involved...

------------------------------------------------------------------------
Sankaran Narayanan                  http://www.cs.columbia.edu/~sankaran

> -----Original Message-----
> From: owner-altq@csl.sony.co.jp [mailto:owner-altq@csl.sony.co.jp]On
> Behalf Of Nicolas Christin
> Sent: Wednesday, October 03, 2001 5:37 PM
> To: ALTQ Mailing List
> Subject: [altq 1066] Accuracy of atobps(const char*)
> 
> 
> 	Hi everyone,
> 
> 	I have a really dumb question. In 
> ALTQ/tbrconfig/tbrconfig.c, one can
> find the following procedure, which is in charge of converting a string
> describing a throughput/rate to an actual number (u_long):
> 
> static u_long
> atobps(const char *s)
> {
>         u_long bandwidth;
>         char *cp;
>                         
>         bandwidth = strtoul(s, &cp, 0);
>         if (cp != NULL) {
>                 if (*cp == 'K' || *cp == 'k')
>                         bandwidth *= 1000;
>                 else if (*cp == 'M' || *cp == 'm')
>                         bandwidth *= 1000000;
>                 else if (*cp == 'G' || *cp == 'g')
>                         bandwidth *= 1000000000;
>         }
>         return (bandwidth);
> }
> 
> 	My question is: Is this procedure correct? I thought 1 Kbps 
> = 2^10 bps
> and *not* 1000 bps (same for Mbps, 2^20 != 10^6, and Gbps). 
> 
> 	It's just a minor detail anyways, but I thought I'd just 
> ask. If I am
> correct, we should have a patch looking like:
> 
> - bandwidth *= 1000;
> + bandwidth <<= 10;
> -------
> - bandwidth *= 1000000;
> + bandwidth <<= 20;
> -------
> - bandwidth *= 1000000000;
> + bandwidth <<= 30;
> 
> 	Cheers,
> 	--Nick
>  
> ------------------------------------------------------------------------
> Nicolas Christin                        nicolas@cs.virginia.edu         
> Multimedia Networks Group               http://www.cs.virginia.edu/~nc2y
> ------------------------------------------------------------------------