Hauke Mehrtens wrote: > I did some more testing with and without the patch. With the patch it > works when the rate is limited to 11M with pid and with minstrel > algorithm. Without this limitation it works only with pid. Of course, when you set a manual rate, the rate-setting method is immaterial. > Without this patch it also works with the limiting the rate to 11M with > both rate control algorithms, but with minstrel it also works without > limiting the rate. Without the patch, minstrel is internally limited to 1 Mb/s TX rate. > Without the patch, without limiting the rate and with minstrel iwconfig > reports a rate of 1Mb/s, but I get 3Mb/s with wget, properly an other > bug. With pid the rate goes up to 54Mb/s and it does not work. The rate reported by iwconfig is for TX. The throughput from wget is dependent on the TX rate of the AP. This info is useful, but the rates from wput, or any other transmitting program, are more instructive to find TX problems. > With the patch, without limiting the rate and with minstrel iwconfig > reports a rate of 24Mb/s and it does not work. With pid it starts with > 1Mb/s and the connection is stable. The rate is changing between 1mb/s, > 2Mb/s, 5.5Mb/s and 9Mb/s. Please submit your rc_stats data for the minstrel case. It is obvious that the RTL8187L does not do as well as other wireless chips in low S/N conditions. It is also possible that some flavors of the chips are better than others. In my Netgear WG111v2 with the patch and using the minstrel algorithm, I get a TX rate between 21 and 22 Mb/s as measured with tcpperf at a distance of 2 m from my AP. The rate reported by iwconfig is 54 Mb/s. When I move to ~15 m from the AP, the rate falls off to ~12 Mb/s with iwconfig reporting 24-48 Mb/s. At 20 m, the connection is essentially unusable - sounds like yours the iwconfig rate is 24 Mb/s. We are clearly flying blind here as we do not have a data sheet for the RTL8187L. I got some help from the Realtek engineer regarding the retry info, but that is the limit of my info. I do have what I think is the latest version of the vendor driver, and I will compare the power setting code with theirs. The patch definitely helps my device as long as the reported signal level is above -60 dBm. With lower signals, then it doesn't work very well. Your device must be in that range. Larry -- To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
