Re: mac80211_hwsim: multiple antennas
On Wed, Nov 21, 2018 at 01:30:40PM +0100, Johannes Berg wrote: > On Wed, 2018-11-21 at 09:26 -0300, Ramon Fontes wrote: > > Sorry. I meant throughput. > > Yeah, well. There's no actual medium (access) simulation in hwsim, so > the actual throughput you get only depends on the performance of your > CPU. 33Mbps seems quite low, but of course I don't know where you're > running this. > > Then again, you say you're using wmediumd, so perhaps that *does* have > some sort of medium simulation these days? I didn't think it has it but > I may very well be wrong on that. wmediumd does simulate the medium, but it only ever handled simple OFDM PHY rates, not MCS or aggregation or anything else from the last decade :) -- Bob Copeland %% https://bobcopeland.com/
Re: mac80211_hwsim: multiple antennas
On Wed, 2018-11-21 at 10:09 -0300, Ramon Fontes wrote: > Yes. I thought that it could be related to the number of antennas (or > something else) after seeing this result > (https://bobcopeland.com/blog/2014/09/wmediumd-speed-test/) provided > by wmediumd. I guess you should instead look at the code referenced by Bob in the next post? https://bobcopeland.com/blog/2014/11/functional-bitrate-sim/ johannes
Re: mac80211_hwsim: multiple antennas
On Thu, 2018-11-22 at 09:18 -0300, Ramon Fontes wrote: > > Then I found the Beacon frame TX rate configuration from hostapd.conf. That's really not relevant for the issue at hand though? johannes
Re: mac80211_hwsim: multiple antennas
I haven't used hwsim with hostapd yet, but I'm pretty sure, that it is possible to change the beacon rate (as we changed them in 802.11s Mesh Mode). I'm think the problem here is, that you can't assign a HT rate for beacons, because action/management frames are always sent with base-rate or at least they need to be sent with legacy-rates in mixed mode. I'm not sure, whether hwsim can use the "green-field" mode, in which everything is sent with ht-rates, but most drivers for real devices do not support green-field mode. So the highest beacon rate seems to be 54.0 mbit/s. Nonetheless, hwsim do not simulate the real timing behavior of this rates. Since hostapd is only responsible for the association protocol but not data transmission, it should not decrease the seen rates extremely, but I have no numbers for it. As more effort you use to simulated the on air behavior, you decrease the throughput rates. kind regards Benjamin Am 22.11.2018 um 13:18 schrieb Ramon Fontes: > I've tested the same setup with adhoc and the throughput seems to be > higher as expected. Thus, the problem comes up when I use hostapd. > > I noticed that the maximum supported rate by hostapd (legacy) is 54Mbps. > Extended supported rates: 24.0 36.0 48.0 54.0 > > Then I tried to modify mac80211_hwsim in order to provide more rates > and it didn't work: > Extended supported rates: 24.0 36.0 48.0 54.0 54.0 54.0 54.0 > > The maximum supported rate is always 54Mbps. > > Then I found the Beacon frame TX rate configuration from hostapd.conf. > After trying to enable the beacon rate (e.g. beacon_rate=ht:1) I got > the following msg: > ap1-wlan1: interface state UNINITIALIZED->HT_SCAN > Using interface ap1-wlan1 with hwaddr 02:00:00:00:02:00 and ssid "ssid-ap1" > nl80211: Driver does not support setting Beacon frame rate (HT) > > and hostapd doesn't work. > > Could this problem be related to mac80211_hwsim? > > Em qua, 21 de nov de 2018 às 10:09, Ramon Fontes > escreveu: >> Yes. I thought that it could be related to the number of antennas (or >> something else) after seeing this result >> (https://bobcopeland.com/blog/2014/09/wmediumd-speed-test/) provided >> by wmediumd. >> Em qua, 21 de nov de 2018 às 09:48, Johannes Berg >> escreveu: >>> On Wed, 2018-11-21 at 09:47 -0300, Ramon Fontes wrote: Okay! Thanks a lot. Indeed, the maximum throughput when using only hwsim is in the order of Gbp/s. >>> So I guess the question is where it's going wrong - is it the wmediumd >>> overhead (netlink), some air simulation in wmediumd, etc. >>> >>> johannes >>> -- M.Sc. Benjamin Beichler Universität Rostock, Fakultät für Informatik und Elektrotechnik Institut für Angewandte Mikroelektronik und Datentechnik University of Rostock, Department of CS and EE Institute of Applied Microelectronics and CE Richard-Wagner-Straße 31 18119 Rostock Deutschland/Germany phone: +49 (0) 381 498 - 7278 email: benjamin.beich...@uni-rostock.de www: http://www.imd.uni-rostock.de/
Re: mac80211_hwsim: multiple antennas
I've tested the same setup with adhoc and the throughput seems to be higher as expected. Thus, the problem comes up when I use hostapd. I noticed that the maximum supported rate by hostapd (legacy) is 54Mbps. Extended supported rates: 24.0 36.0 48.0 54.0 Then I tried to modify mac80211_hwsim in order to provide more rates and it didn't work: Extended supported rates: 24.0 36.0 48.0 54.0 54.0 54.0 54.0 The maximum supported rate is always 54Mbps. Then I found the Beacon frame TX rate configuration from hostapd.conf. After trying to enable the beacon rate (e.g. beacon_rate=ht:1) I got the following msg: ap1-wlan1: interface state UNINITIALIZED->HT_SCAN Using interface ap1-wlan1 with hwaddr 02:00:00:00:02:00 and ssid "ssid-ap1" nl80211: Driver does not support setting Beacon frame rate (HT) and hostapd doesn't work. Could this problem be related to mac80211_hwsim? Em qua, 21 de nov de 2018 às 10:09, Ramon Fontes escreveu: > > Yes. I thought that it could be related to the number of antennas (or > something else) after seeing this result > (https://bobcopeland.com/blog/2014/09/wmediumd-speed-test/) provided > by wmediumd. > Em qua, 21 de nov de 2018 às 09:48, Johannes Berg > escreveu: > > > > On Wed, 2018-11-21 at 09:47 -0300, Ramon Fontes wrote: > > > Okay! Thanks a lot. > > > > > > Indeed, the maximum throughput when using only hwsim is in the order of > > > Gbp/s. > > > > So I guess the question is where it's going wrong - is it the wmediumd > > overhead (netlink), some air simulation in wmediumd, etc. > > > > johannes > >
Re: mac80211_hwsim: multiple antennas
Yes. I thought that it could be related to the number of antennas (or something else) after seeing this result (https://bobcopeland.com/blog/2014/09/wmediumd-speed-test/) provided by wmediumd. Em qua, 21 de nov de 2018 às 09:48, Johannes Berg escreveu: > > On Wed, 2018-11-21 at 09:47 -0300, Ramon Fontes wrote: > > Okay! Thanks a lot. > > > > Indeed, the maximum throughput when using only hwsim is in the order of > > Gbp/s. > > So I guess the question is where it's going wrong - is it the wmediumd > overhead (netlink), some air simulation in wmediumd, etc. > > johannes >
Re: mac80211_hwsim: multiple antennas
On Wed, 2018-11-21 at 09:47 -0300, Ramon Fontes wrote: > Okay! Thanks a lot. > > Indeed, the maximum throughput when using only hwsim is in the order of Gbp/s. So I guess the question is where it's going wrong - is it the wmediumd overhead (netlink), some air simulation in wmediumd, etc. johannes
Re: mac80211_hwsim: multiple antennas
Okay! Thanks a lot. Indeed, the maximum throughput when using only hwsim is in the order of Gbp/s. Em qua, 21 de nov de 2018 às 09:30, Johannes Berg escreveu: > > On Wed, 2018-11-21 at 09:26 -0300, Ramon Fontes wrote: > > Sorry. I meant throughput. > > Yeah, well. There's no actual medium (access) simulation in hwsim, so > the actual throughput you get only depends on the performance of your > CPU. 33Mbps seems quite low, but of course I don't know where you're > running this. > > Then again, you say you're using wmediumd, so perhaps that *does* have > some sort of medium simulation these days? I didn't think it has it but > I may very well be wrong on that. > > Try pure hwsim (no wmediumd) and see if you get gigabits, if not then > your system is probably just too slow. > > And then I guess you'd have to look at wmediumd in more detail to see if > it has some simulations here? I don't know. > > johannes >
Re: mac80211_hwsim: multiple antennas
On Wed, 2018-11-21 at 09:26 -0300, Ramon Fontes wrote: > Sorry. I meant throughput. Yeah, well. There's no actual medium (access) simulation in hwsim, so the actual throughput you get only depends on the performance of your CPU. 33Mbps seems quite low, but of course I don't know where you're running this. Then again, you say you're using wmediumd, so perhaps that *does* have some sort of medium simulation these days? I didn't think it has it but I may very well be wrong on that. Try pure hwsim (no wmediumd) and see if you get gigabits, if not then your system is probably just too slow. And then I guess you'd have to look at wmediumd in more detail to see if it has some simulations here? I don't know. johannes
Re: mac80211_hwsim: multiple antennas
Sorry. I meant throughput. Let's considering GI=800ns - 20MHz: Without setting the MCS value the throughput is around 18Mbps. tx bitrate 58.5 MBit/s MCS 6 rx bitrate 48.0 MBit/s Then I set the bitrate to ht-mcs-2.4 13, 14 and finally 15. Although the tx bitrate increases as expected the maximum throughput is around 33Mbps tx bitrate: 117.0 MBit/s MCS 14 rx bitrate: 54.0 MBit/s I also tried the same with 40Mhz and the throughput doesn't change. The maximum throughput is always ~33Mbps. Even when the tx bitrate is 300Mbps the maximum throughput is still ~33Mbps. I noticed that the rx birate is related to the hwsim_rates from mac80211_hwsim, however, increasing the value provided by hwsim_rates doesn't change the throughput. My topology consists of three nodes. 2 clients and one AP working with hostapd. The hostapd config file includes ieee80211n=1 and ht_capab='[HT40+]'. Sometimes I can see that tx bitrate is working at 40Mhz. However, it doesn't work at 40Mhz when there are data traffic. The RSSI of both clients is -60dBm, but I also tried with -50dBm and -40dBm. Tested with both 2.4 and 5Ghz. This is my per.c file (https://github.com/ramonfontes/wmediumd/blob/ieee80211n/wmediumd/per.c). I only check if the rate-flags == IEEE80211_TX_RC_MCS and I force some of the rateset added in my per.c. Given this brief explanation I was wondering if it could be related to the number of antennas. Thank you in advance. Ramon Em qua, 21 de nov de 2018 às 06:11, Johannes Berg escreveu: > > On Tue, 2018-11-20 at 16:33 -0300, Ramon Fontes wrote: > > Would be possible to simulate (or even add support to) multiple > > antennas? I was trying to extend wmediumd to 802.11n according the > > rates provided by http://mcsindex.com/, but the maximum rate I get is > > ~33Mbps, even if I set the MCS index to 15. Given the result, I > > thought that it could be related to the number of antennas. Is my > > assumption correct? > > I'm not sure what you mean by "the maximum rate I get is ~33Mbps". > > Are you trying to actually _measure_ it? Then, on a decent platform, you > should probably get many *gigabits*. > > Or are you trying to look at the rate in "iw station dump" or in > radiotap? Then you should see higher rates since hwsim is configured to > pretend to have 4 spatial streams (IIRC) > > johannes >
mac80211_hwsim: multiple antennas
Would be possible to simulate (or even add support to) multiple antennas? I was trying to extend wmediumd to 802.11n according the rates provided by http://mcsindex.com/, but the maximum rate I get is ~33Mbps, even if I set the MCS index to 15. Given the result, I thought that it could be related to the number of antennas. Is my assumption correct? regards, Ramon