search

seemoo-lab / nexmon_csi

Channel State Information Extraction on Various Broadcom Wi-Fi Chips
313 stars 122 forks source link

ASUS with nexmon frequently crash and reboot when run tcpdump #123

Open yujianyuanhaha opened 4 years ago

yujianyuanhaha commented 4 years ago

hi,

thanks for your amazing teamwork of nexmon, I have been use the tool for a while and get the CSI I want. However, it is annoying to me the router (model ASUS RT-AC86U) will frequently restart when I run tcpdump to collect CSI for longer time. Each time the router reboot I have to flash in nexmon again. It is bit random how long the tcpdump will last, usually cannot last longer than 10min.

Several point I checked

  1. close wifi and other network connection, the host machine is linked to sniffer router only.

  2. tcpdump for shorter time. The router has near 50M internal storage, and can last at most 80 seconds. However, it happen still sometime. If we choose to save .pcap file in external storage /mnt/sda1/ (a USB flash driver), it usually crash down within 10minutes. There is indeed few time it can last as long as 30 miniutes and I got over 1G .pcap files.

This bug has limit my application of long-time data collecting more than 10min.

Anyone meet the issue before and have a solution. @schmittner @DanielAW @jlinktu @matthiasseemoo

Best.

zel602 commented 4 years ago

I meet this issue before，if I I have a period of time between twice tcpdump ,I will get 0 packets all the time,I have to reboot ASUS RT-AC86U,so I keep running tcpdump even if I am moving the position of the transmitter, this may mitigate this problem. By the way,do you meet the phenomenon that the CSI's subcarrier -65 and subcarrier -64 are inconsistent ,when using the 80MHz channel.

yujianyuanhaha commented 4 years ago

@zel602 My problem is ssh break down, not get 0 packet.

For you issue, you may need to open two terminal, in one terminal run tcpdump, while it is stuck in another terminal switch off-n-on interface by wl radio off; wl radio on; ifconfig eth6 off; ifconfig eth6 on, then you can see tcpdump resume collection.

it make sense subcarrier inconsistent since 80MHz bandwidth is compose of 4 20MHz bandwidth, besides null/ guard subcarrier, there should always be sudden inconsistent at each 1/4 point (i.e. -64 0, +64)

zel602 commented 4 years ago

@zel602 My problem is ssh break down, not get 0 packet.

For you issue, you may need to open two terminal, in one terminal run tcpdump, while it is stuck in another terminal switch off-n-on interface by wl radio off; wl radio on; ifconfig eth6 off; ifconfig eth6 on, then you can see tcpdump resume collection.

it make sense subcarrier inconsistent since 80MHz bandwidth is compose of 4 20MHz bandwidth, besides null/ guard subcarrier, there should always be sudden inconsistent at each 1/4 point (i.e. -64 0, +64)

Thanks for your answer!You have solved my another question! For your issue,I think you can change this parameter in router setting,which may be helpful.

router

By the way,for the subcarrier inconsistent problem,what makes me feel confused is the CSI's subcarrier -65 and subcarrier -64 are inconsistent, however,the phase of subcarrier 63 and subcarrier 64 are consistent.Do you meet same phenomenon before?How much bandwith between subcarrier -65 and subcarrier -64 ?

channel63_64

CanI get your contact information?My Email is 419046121@qq.com.Maybe we can have more communication.

yujianyuanhaha commented 4 years ago

hi @zel602

I disable idle timeout very early, it should be other reason, I feel like it maybe memory overflow.

Again, since there is a gap between each 20M bandwidth, it is not surprise there is sudden change/ inconsistent between -65,-64. I am not clear of the size of gap as well. A easier way is drop few subcarriers since CSI is pretty rich.

Feel free to post your question right here, this is the power of open source and Github.

zel602 commented 4 years ago

hi @zel602

I disable idle timeout very early, it should be other reason, I feel like it maybe memory overflow.

Again, since there is a gap between each 20M bandwidth, it is not surprise there is sudden change/ inconsistent between -65,-64. I am not clear of the size of gap as well. A easier way is drop few subcarriers since CSI is pretty rich.

Feel free to post your question right here, this is the power of open source and Github.

Hi!

I never meet the situation ssh break down.May the firmware you use is the matter.

Could you share the phase figure for your measured CSI in 80MHz?I think we can get some results by comparing the phase figure.

yujianyuanhaha commented 4 years ago

several reason I feel like it may ease the crash down:

  1. change the router Ip from 192.168.0.x to 192.168.2.x
  2. save into smaller .pcap file
  3. load the nexmon tool via /sbin/rmmod dhd; /sbin/insmod /jffs/dhd.ko directly in the router
tweigel-dev commented 3 years ago

sry for the trivial question but how you install tcpdump at the Asus, because my asus isn't conneted to the internet. To copy the tcp bin does not work for me. @yujianyuanhaha and @zel602

zeroby0 commented 3 years ago

Install TCPdump

All the below commands should be run on your computer; as signified by the user@vm$. The compile toolchain and environment is setup in 8th and 9th steps of https://github.com/seemoo-lab/nexmon_csi#bcm4366c0. You would have run those steps while installing nexutil. Run these commands in the same terminal, or run 9th command before running these scripts in a new terminal.

I connect my Asus's wan port to my internet routers lan port and assign a static IP to the Asus. Then my laptop is connected to the main router via WiFi. So Asus has internet, and I can move around my house without having to drag the Asus around. I'm working on a user friendly guide to Nexmon_csi for rt-ac86u like my Pi version :)

tweigel-dev commented 3 years ago

Thanks for the tutorial. but one step failed while Cross compile LibPCAP:

 CC=aarch64-linux-gcc AR=aarch64-linux-ar RANLIB=aarch64-linux-ranlib AS=aarch64-linux-as LD=aarch64-linux-ld NM=aarch64-linux-nm ./configure --host=x86_64-unknown-linux-gnu --with-pcap=linux
configure: WARNING: If you wanted to set the --build type, don't use --host.
    If a cross compiler is detected then cross compile mode will be used.
checking build system type... x86_64-unknown-linux-gnu
checking host system type... x86_64-unknown-linux-gnu
checking target system type... x86_64-unknown-linux-gnu
checking for x86_64-unknown-linux-gnu-gcc... aarch64-linux-gcc
checking for C compiler default output file name... 
configure: error: C compiler cannot create executables
See `config.log' for more details.

@zeroby0 Did you have a clue whats going on ? Also which makefile you mean to run make on ?

zeroby0 commented 3 years ago

I saw that error yesterday in one of the issues. Have you installed g++ and ran step 8 and 9? I think it is because of step 9.

Edit: found it: https://github.com/seemoo-lab/nexmon_csi/issues/34#issuecomment-577068333

crownz-sec commented 2 years ago

several reason I feel like it may ease the crash down:

  1. change the router Ip from 192.168.0.x to 192.168.2.x
  2. save into smaller .pcap file
  3. load the nexmon tool via /sbin/rmmod dhd; /sbin/insmod /jffs/dhd.ko directly in the router

Hi, have you solved the reboot problem yet? I found the same reboot issue when using UbiLoate based on Nexmon. So, I came to Nexmon project to find the solution. I'll try what you said, but I still want to ask if there are other methods, and ultimately how long can you keep the router from rebooting? I'm sorry that I'm responding to the message from a long time ago, but this issue really bothers me so much that I have to do this.