On 4/2/07, Gerrit Renker <gerrit@xxxxxxxxxxxxxx> wrote:
Quoting Ian McDonald: | > +/** | > + * Macro for exponentially weighted moving average | > + * @weight: Weight to be used as damping factor, in units of 1/10 | > + * Beware that @val is evaluated multiple times. | > + */ | > +#define TFRC_EWMA(val, newval, weight) \ | > + val = val? ((weight) * val + (10 - (weight)) * (newval)) / 10 : (newval) | > + | Just wondering why you pass weight when you only use 9 anyway and it | just adds a step. Is it used in following patches with a different | weight? | In the following patches it is used one more time, which is for the `Preventing Oscillations' changeset (this is not in the next patch set but in the one after that); here again the weight 9/10 is used. I think it is useful to keep the `weight' argument, since RFC 3448 suggests to use a different weight for the RTT sample when the `Oscillation Prevention' e.g. is not deployed (a q close to zero). I think it is better to keep the weight argument, it allows changing the code later, while still providing the abstraction: for CCID 3 there are 4 uses of this macro alone.
OK. Acked-by: Ian McDonald <ian.mcdonald@xxxxxxxxxxx> -- Web: http://wand.net.nz/~iam4/ Blog: http://iansblog.jandi.co.nz WAND Network Research Group - To unsubscribe from this list: send the line "unsubscribe dccp" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
