On Thu, 19 Jan 2006 17:58:13 -0600 Saurabh Jain <[EMAIL PROTECTED]> wrote:
> Hi All, > > After going through the TCP code in linux kernel it seems that TCP in > linux stores srtt in milliseconds and it is inflated by a factor of 8. It is clock ticks scaled by 8. The clock ticks are dependent on the HZ configured (default is 250hz) now. > Does this means that TCP in linux can measure any rtt which is greater > than 125 microseconds? The code will store a value of 1 for srtt > variable when the rtt is 125 micro seconds (125 microsecond * 8 = 1 > millisecond). The measurement is limited by the resolution of the system clock HZ (see above). So with 250hz the minimum value is 4ms. The scaling is just to make the average calculation faster (see old Van Jacobsen code in 4.2 BSD). - To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
