IRC channel logs
2024-11-30.log
back to list of logs
<damo22>youpi: im reading the libbpf code but i cant seem to find a function that sends a network packet <damo22>i think i need a SOCK_DGRAM not a SOCK_RAW too because libpcap uses that <damo22>i understand that opening the socket allows me to call rump_sys_ioctl(socket, ...) with different functions to configure the net device, but i cant seem to find how to actually send/rx a packet <youpi>sneek later tell gfleury sure, it's in my queue. I was looking at failed logs, next target could be pthread_sigmask and pthread_set/getspecific <sneek>youpi, gfleury says: few days ago i have sent a pachset for pthread_contrattr_*. If you have a time this weekend take a look at it <youpi>damo22: the tcpdump code will show you how to receive packets <youpi>the ping code can tell you how to send a packet <damo22>well, i think it says device_read and device_write, but i dont know how to implement these to the NIC via rump, i seem to be missing some docs for rump <damo22>i imagine it should be easy like rump_sys_read(some_fd, ... <youpi>just like tcpdump and ping use it <damo22>/hurd/pfinet: ../../pfinet/ethernet.c:386: setup_ethernet_device: Assertion 'netstat.header_format == 1' failed. <damo22>i guess i need to implement device_get_status and device_set_status properly <damo22>fsysopts: /servers/socket/2: -i /dev/wm0 -a 0.0.0.0 -m 255.0.0.0: Error in unknown error system: FFFFF63C <damo22>youpi: how does pfinet configure the network device? eg address and netmask, it doesnt seem to make sense how it implements the ioctl itself <damo22>surely netdde configures the eth0 <damo22>is it a bad idea to try to use rumpnet + pfinet? <gnu_srs1>Is this a bug in dash? [ -x `command -v restorecon` ] && echo "\$?="$? returns $?=0 not 1?? <youpi>damo22: pfinet does not need to tell the device about ip, it's all software, the ethernet card doesn't care <youpi>rumpnet+pfinet fully makes sense (rumpnet+lwip too) <youpi>pfinet or lwip just create ethernet frames and give it to rumpnet, and gets ethernet frames from it <youpi>btw, now I realize: it's not ping that you'd want to look at for sending frames, since it creates only icmp/ip packets, it's rather dhcp or arping (which create ethernet frames) <youpi>gnu_srs1: $? only provides the return value of the last whole shell command , not of the last piece of command <youpi>as well as true ; true && echo $? <gnu_srs1> `command -v restorecon` returns an empty string <youpi>you're just missing "" to get what you want <youpi>so it's indeed seen by -x as empty string <youpi>so [ -x "`command -v restorecon`" ] && echo "\$?="$? <youpi>you really need to think by logic, not by intuition <gnu_srs1>OK, so I have to embed into parenthesis? <youpi>no, I didn't write any parentheses <gnu_srs1>Sorry, I meant qoutes, or whatever "" is called <gnu_srs1>Found it: ' = apostrophe, " = quotation mark, ` = backtick/grave accent <gnucode>heyo, I'm updating my T43 and the "change-logs" / "messages" just said, that next week apt will no longer use "apt-key" to verify signatures. <sneek>gnucode, you have 1 message! <sneek>gnucode, ZhaoM says: I successfully install luakit in the Hurd on my T60, thanks for reminding me some work is needed <gnucode>Apparently it won't be there next week. Are we ready for that change? <gnucode>sneek later tell ZhaoM Nicely done! How well does luakit work on the Hurd? <gnucode>sneek later tell ZhaoM nice job on the rtc by the way! That was a very quick 1st contribution! <sneek>gfleury, you have 1 message! <sneek>gfleury, youpi says: sure, it's in my queue. I was looking at failed logs, next target could be pthread_sigmask and pthread_set/getspecific