**** BEGIN LOGGING AT Wed Jun 20 03:00:08 2018 Jun 20 04:35:48 I've compiled flex-2.6.0 and bison-3.0.4 for OpenWRT MIPS Jun 20 04:36:17 The trick was setting up ./configure properly (-host=mipsel-openwrt-linux) as well as initializing PATH and STAGING_DIR variables Jun 20 06:00:26 hey, is next release coming by the end of this month, right? Jun 20 06:03:35 airwind: it'll come when it's ready. but that was the goal last i heard Jun 20 06:05:58 24th for a rc1 release Jun 20 06:08:47 blogic: thanks for the better answer :-) Jun 20 06:16:24 stintel: i'm leaving for 3 days Jun 20 06:16:29 stintel: ping me on weekend/Monday please Jun 20 06:22:18 neoraider: am I right that you have access to a mt7621 DIR-860L B1? would you mind to test an image with https://github.com/openwrt/openwrt/pull/925 applied? Jun 20 06:23:07 neoraider: I would like to ensure that the changes to the lzma loader doesn't break the relocate kernel used for the DIR-860L B1 Jun 20 06:40:54 I hate the "when it's ready" answer Jun 20 06:41:38 airwind: sure, but you would hate a release that doesn't work even more Jun 20 07:09:47 airwind: hate is a very hard word to describe the voluntary work of others ;) Jun 20 07:10:18 airwind: you are however invited to invest 20 hours of your free time, weekly over 10 years to help us Jun 20 07:12:46 Dude, just drop it. I have this tendency to dislike relative answers when clearly absolute ones are expected... i.e. "when it's done" vs "DDMMYYYY" Jun 20 07:13:04 seen that so many times... Jun 20 07:13:34 airwind: well its like self harming then Jun 20 07:13:46 hey everyone, question regarding hotplug scripts for 00_wwan.sh Jun 20 07:13:47 you know there is no answer, then you ask regardless just to complain Jun 20 07:13:51 dont get the point Jun 20 07:13:59 its like pouring oil into fire Jun 20 07:14:10 landwork: go for it Jun 20 07:14:23 are they supposed to actually call functions in /lib/netifd/netifd-proto.sh? like proto_set_available ? Jun 20 07:14:33 i was debugging the scripts and found that they don't. Jun 20 07:15:51 it should Jun 20 07:16:04 to cut a long storey short i'm trying to get wwan proto working with two types of cellular engines… one exposes the tty devices as serial option drivers (ttyUSB), and the other (newer) uses CDC ultimately (ttyACM)… Jun 20 07:16:05 it'll tell netifd that the device appeared and it can bring up the interface Jun 20 07:16:36 should the wwan proto not do that for you already ? Jun 20 07:16:46 no… Jun 20 07:16:58 when I hotplug between modem types it's not working. Jun 20 07:17:01 ah for the same vid:pid pair i guess Jun 20 07:17:14 there is another issues with ttyACM… but I have fixed this. Jun 20 07:17:17 then ran into this problem. Jun 20 07:17:26 show me your uci section please ? Jun 20 07:17:32 is there a device in there ? Jun 20 07:17:43 the uci section doesn't have the device configuration. Jun 20 07:17:48 wwan should set it implictly based on that json file index we have Jun 20 07:18:00 weird Jun 20 07:18:08 the wwan proto should use the vid:pid files to point to the usb endpoints. Jun 20 07:18:11 and that is not working Jun 20 07:18:25 and your /lib/network/wwan/$vid:$pid file exist Jun 20 07:18:34 yes i created them Jun 20 07:18:36 so in /lib/netifd/proto/wwan.sh... Jun 20 07:18:57 add "logger foo" lines to the hotplug script and check if they appear in the syslog Jun 20 07:18:59 the first problem i had was the line where it interrogates the tty devices via the usb endpoint Jun 20 07:19:06 I had to modifiy it to this -> ttys=$(ls -d /sys/bus/usb/devices/$devicename/${devicename}*/tty?* /sys/bus/usb/devices/$devicename/${devicename}*/tty/tty?* | sed "s/.*\///g" | tr "\n" " ") Jun 20 07:19:28 the current code does not take into account that ttyUSB and ttyACM are represented differently in the sys fs Jun 20 07:19:40 ttyACM has another subdirectory "tty" Jun 20 07:19:44 which i fixed above. Jun 20 07:19:46 that's the first problem. Jun 20 07:20:02 after fixing that it works with ttyACM style USB-CDC devices. Jun 20 07:20:16 yes, they don't occur. Jun 20 07:20:22 logging lines that is. Jun 20 07:20:53 landwork: ok, this stuff used to work when i built it but not touched it in years Jun 20 07:20:56 :-) Jun 20 07:20:58 ermmm Jun 20 07:20:58 yeah :) Jun 20 07:21:02 that's cool i'm happy to fix it! Jun 20 07:21:12 I just don't want to break anything, so I'll definitely submit a patch Jun 20 07:21:16 i think things changed over the years. Jun 20 07:21:19 :P Jun 20 07:21:31 so first problem aside… the second issues. Jun 20 07:21:56 at the top of /etc/hotplug.d/usb/00_wwan.sh if I add ". /lib/netifd/netifd-proto.sh" it looks to work correctly Jun 20 07:22:05 * blogic builds an image with wwan inside Jun 20 07:22:13 i think it works fine if you explicitly define the device in UCI... Jun 20 07:22:15 but I don't want to do that. Jun 20 07:22:39 I want it to determine it automagically with the /lib/network/wwan/vid:pid files... Jun 20 07:23:39 yeah Jun 20 07:23:48 the whole point of the wwan proto is that you dont need to do so Jun 20 07:23:57 yeah i just confirmed it. Jun 20 07:24:03 exactly :) Jun 20 07:24:08 it's actually pretty awesome Jun 20 07:24:24 we've been using 3g proto for a while and had issues with the ttyUSB device names bouncing around Jun 20 07:24:29 this wwan approach will fix it. Jun 20 07:24:35 just need to fix a couple of bugs. Jun 20 07:24:51 so I can do this in the next couple of days, is it ok to submit patches still through github ;P Jun 20 07:24:57 pull requests that is. Jun 20 08:31:00 mkresin: regarding your question @ 08:22:18, hexa also has a dir-860l Jun 20 09:26:05 dangole ping Jun 20 09:27:58 hey everyone, I just noticed something strange while doing some tests: I have a wifi interface bridged to the LAN, I'm testing download and upload traffic monitoring. Jun 20 09:27:58 I assumed RX stands for received, so download, while TX stands for transmitted so upload, this works when monitoring the LAN interface, but when monitoring the bridged wifi it seems to be reversed, I keep on downloading stuff and TX (upload) is roughly always double the RX (download), anyone can hint why is that? Jun 20 09:56:37 hexa: 08:22 < mkresin> neoraider: am I right that you have access to a mt7621 DIR-860L B1? would you mind to test an image with https://github.com/openwrt/openwrt/pull/925 applied? Jun 20 09:56:41 hexa: 08:23 < mkresin> neoraider: I would like to ensure that the changes to the lzma loader doesn't break the relocate kernel used for the DIR-860L B1 Jun 20 09:56:52 rotanid: thanks a lot for the pointer Jun 20 10:50:39 Hello Jun 20 10:51:12 Is TP-WR1043ND v2 supported? On the site listing I see only v2.1 Jun 20 10:51:27 But the firmware is for v2.x Jun 20 10:54:28 baby_tsunami: the numbers behind the dot are usually irrelevant Jun 20 12:30:04 hi, is there a way to build openwrt and then move around the openwrt folder and build again? Jun 20 12:31:22 do you have any particular problem? Jun 20 12:31:36 do you just mean mv openwrt-root some-new-path && cd newpath && make? Jun 20 12:31:44 that "just works" as far as I know. Jun 20 12:33:01 mwarning: this is not really supported Jun 20 12:33:33 mwarning: most autoconf / automake / libtool based projects will fail hard since those build systems embed absolute path references all over the build tree Jun 20 12:33:35 I thought all the scripðts exported staging dir at the start? do things get hard paths baked into them? Jun 20 12:33:59 bah, autofail again Jun 20 12:34:02 the openwrt buildroot tools itself might work (they're patched for relocatability for the SDK) Jun 20 12:34:32 but you will need to perform a "make dirclean" after moving the buildroot folder which I guess contradicts whatever you attempt to achieve with that Jun 20 12:43:01 jow: alright, thanks! Jun 20 12:43:47 karlp: I want to create a portable system to build openwrt packages Jun 20 12:44:07 hm, maybe a chroot might do the trick.. Jun 20 12:45:08 hey jow Jun 20 12:45:16 mwarning: that exists, its called SDK Jun 20 12:45:31 mwarning: openwrt SDK is a portable packge building thingy Jun 20 12:45:40 heya peepz Jun 20 12:45:48 'lo Jun 20 12:45:56 stintel: y0 Jun 20 12:45:59 mwarning: there's a lot of documentation on using the sdk on the wiki too. Jun 20 12:46:13 ok, I will checkout the sdk Jun 20 12:46:16 can you tell me if I've done this right or not? (pretty much copy/paste): https://github.com/joeholden/fw3/commit/afc95a95a202006d2ee4ba392f52e0f2a5dbd92b Jun 20 12:46:17 I think my lirc package is about ready to go into the packages feed Jun 20 12:46:19 :D Jun 20 12:47:36 can I get a quick review for http://git.openwrt.org/5a31a627 - it's just a oneliner Jun 20 12:48:17 huh > Jun 20 12:48:24 looks like a ath10k bug Jun 20 12:48:35 ath10k should not use the symbnol if its not defined Jun 20 12:49:01 stintel: it looks like a work around, hiding the problem but i would need to dive into the ath10k code Jun 20 12:49:33 blogic: the problem is that ath10k selects kmod-thermal which makes that symbol visible Jun 20 12:49:40 blogic: it was added in 4.12 Jun 20 12:50:08 blogic: other targets have CONFIG_THERMAL and friends enabled in their config-4.14 Jun 20 12:50:14 blogic: but ath79 doesn't Jun 20 12:53:54 ah ok Jun 20 12:53:58 then go for it Jun 20 12:54:52 aight, danke :) Jun 20 12:56:17 so if I have a kmod package, do I need to do anything special to have kmodloader load it by default at boot? Jun 20 12:56:26 I think so Jun 20 12:56:41 AUTOLOAD:=$(call AutoProbe,lirc-rpi) Jun 20 12:56:43 e.g. Jun 20 12:56:46 aaah Jun 20 12:56:50 thanks Jun 20 12:57:58 jwh: what is netflow? Jun 20 12:59:11 flow export (of traffic) Jun 20 12:59:26 I managed to get that ipt-netflow thing woring Jun 20 12:59:29 working Jun 20 13:00:45 seems to work reasonably well, gonna be faster than any of the existing solutions in the packages tree anyway Jun 20 13:01:19 blogic: in the PR for c59v2 you added the "needs changes" label without writing which changes? https://github.com/openwrt/openwrt/pull/1054 Jun 20 13:01:54 the only one I can see is softflowd, which sets interfaces to promisc and captures everything, seems a bit silly when conntrack already has the info Jun 20 13:04:39 I'm not sure of the ordering though, it might need to be before offload (if enabled), I'm not netfilter expert Jun 20 13:04:43 no* Jun 20 13:05:37 blogic: I think we need to mark the iptables package as nonshared Jun 20 13:05:45 https://www.mail-archive.com/source-changes@openbsd.org/msg99141.html Jun 20 13:06:33 blogic: since flow offloading is only available on kernel 4.14+ while iptables is shared among 4.14 and 4.9 targets, many offload capable-targets end up with an iptables built without offloading support Jun 20 13:06:55 I think this is the issue that PKG_RELEASE bump patch meant to fix Jun 20 13:13:08 can the SDK be used to build images as well? Jun 20 13:16:59 dangole ping2 Jun 20 13:17:24 blogic: also, ~2 weeks ago you said the C7v5 PR looks good and wanted to take a closer look before merging ;) https://github.com/openwrt/openwrt/pull/893 Jun 20 13:23:04 jow you have a final decision regarding golang in 18.06 I'll close the PR if it's considered "bloating churn" https://github.com/openwrt/packages/pull/6314 Jun 20 13:26:17 jow: so, any thughts? am I misuderstanding it? Jun 20 13:31:16 Widely used D-Link modem/router under mass attack by potent IoT botnet Jun 20 13:31:21 https://arstechnica.com/information-technology/2018/06/widely-used-d-link-modemrouter-under-mass-attack-by-potent-iot-botnet/ Jun 20 13:32:14 Any one know if that router is one that uses openwrt as the fermwair Jun 20 13:32:55 mkresin: I'll check Jun 20 13:33:00 rotanid: and then i went on vacation Jun 20 13:33:02 Do D-Link even use openwrt as a base or is that netgair Jun 20 13:33:16 rotanid: your fault for using github. use email and i'll merge it really quickly Jun 20 13:33:29 github is just so shitty things simply get lost amongst the piles of poo poo Jun 20 13:33:50 blogic: i know and it's not my PR so not my fault ;D Jun 20 13:34:02 ah i recall :-) Jun 20 13:34:16 aparcar: merged Jun 20 13:34:19 you said we're allowed to nag you from time to time Jun 20 13:34:24 sure Jun 20 13:34:29 tell me the link please Jun 20 13:34:53 i already did a few minutes ago? Jun 20 13:35:21 ,issed it Jun 20 13:35:56 jow thanks! Jun 20 13:35:57 15:01:19 < rotanid > blogic: in the PR for c59v2 you added the "needs changes" label without writing which changes? https://github.com/openwrt/openwrt/pull/1054 Jun 20 13:36:00 15:17:24 < rotanid > blogic: also, ~2 weeks ago you said the C7v5 PR looks good and wanted to take a closer look before merging ;) https://github.com/openwrt/openwrt/pull/893 Jun 20 13:36:54 blogic: mkresin and whoelse was involved in the discussion few days ago, i found a forum discussion on the github versus mail/patchwork thingy: https://forum.lede-project.org/t/github-pr-vs-mailing-list-patchset/15657 Jun 20 13:37:06 rotanid: adrian request spaces vs tabs to be fixed Jun 20 13:37:27 he wrote "I'm not in a position to tell what's preferred or not" Jun 20 13:37:37 and he's no maintainer, so i wasn't sure what to expect Jun 20 13:38:13 he complained about a valid point Jun 20 13:38:23 there is no hierachy or one voice has more say than another thing Jun 20 13:38:28 yeah, looking at it that's valid Jun 20 13:38:33 this is post lede openwrt and not openwrt Jun 20 13:38:37 :D Jun 20 13:39:44 Tapper: interesting, dlink is produced by cameo afaik and they use vendor SDKs + shitty own code Jun 20 13:39:51 Tapper: try looking for a GPL drop Jun 20 13:44:03 blogic: thanks again Jun 20 13:47:54 build #50 of gemini/generic is complete: Failure [failed images] Build details are at http://phase1.builds.lede-project.org/builders/gemini%2Fgeneric/builds/50 blamelist: Johann Neuhauser , Hans Dedecker , Daniel Golle , INAGAKI Hiroshi , Mathias Kresin Jun 20 13:47:54 Jun 20 13:50:18 blogic: Appears to be a Broadcrap Jun 20 13:51:14 * ldir notices Monkeh still cannot spell Jun 20 13:52:10 lol Jun 20 13:52:12 Monkey that is, not Broadcrap :-) Jun 20 13:53:16 :) Jun 20 13:59:13 Yay, patchwork is fast today. Let's see how that turns out. Jun 20 14:19:35 hexa: thanks for the testing Jun 20 14:19:44 sure Jun 20 14:28:41 I still have problems using the openwrt sdk: https://paste.ubuntu.com/p/qYyJFjfzBY/ Jun 20 14:30:07 the package in question exists Jun 20 14:30:12 mwarning: maybe your hsot system is too new Jun 20 14:30:23 the host is new.. Jun 20 14:30:30 arch linux... Jun 20 14:30:33 or glibc invented clever new tricks to not work Jun 20 14:30:41 probably both Jun 20 14:30:48 hm speaking of which Jun 20 14:30:55 I should probably file a bug report Jun 20 14:31:04 gdb seems to be broken using gcc 8 Jun 20 14:31:10 (on the host) Jun 20 14:31:10 don't use gcc 8 then :) Jun 20 14:31:21 builder has gcc 8 :P Jun 20 14:31:38 I broke my alpine builder :( Jun 20 14:31:47 mwarning: what ahppens when you execute ./staging_dir/host/bin/sed ? Jun 20 14:32:03 .sed.bin: loadlocale.c:129: _nl_intern_locale_data: Assertion `cnt < (sizeof (_nl_value_type_LC_TIME) / sizeof (_nl_value_type_LC_TIME[0]))' failed. Jun 20 14:32:04 Aborted (core dumped) Jun 20 14:32:08 xD Jun 20 14:32:28 thats your problem Jun 20 14:32:47 what host os ? Jun 20 14:32:54 arch linux Jun 20 14:32:57 probably need to kick ld.so some more to not do crap Jun 20 14:33:10 mwarning: glibc version? Jun 20 14:33:36 2.27-3 Jun 20 14:33:38 mwarning: ah, can you please try this: Jun 20 14:33:41 export LC_ALL=C Jun 20 14:33:44 ./staging_dir/host/bin/sed Jun 20 14:33:53 does it print a help instead of coredumping then? Jun 20 14:33:59 looks good! Jun 20 14:34:05 cool, so there's your workaround Jun 20 14:34:11 export LC_ALL=C Jun 20 14:34:13 ^^ Jun 20 14:34:17 then do the usual make stuff Jun 20 14:34:37 jow: so... back to the fw3 thing, am I doing it right? (and can you maybe commit it?), or is there a more sensible way of extending it Jun 20 14:34:40 (Anyway, I had to download tools/cmake when building openwrt from the ground up) Jun 20 14:35:57 jwh: can you cpmpare the fw3 executable size before/after ? Jun 20 14:36:15 yeah sure Jun 20 14:38:20 I need to double check the ordering again, may need to come before -j FLOWOFFLOAD Jun 20 14:42:51 meh, 672bytes on x86_64 musl, will check other targets Jun 20 14:43:07 seems excessive Jun 20 14:47:39 mabybe roll a single function that checks if /sys/modules/X/refcnt exists to avoid duplication Jun 20 14:48:53 weird. sysupgrde -v -n /tmp/newimage, and I still have a symlink in /usr/lib/lua/luci/controller from before hand. (only now it's dnagling of course) Jun 20 14:50:12 how does that happen?! Jun 20 14:50:18 makes no sense. done it twice now. Jun 20 14:51:02 ysupgrade abort[ 1446.057564] Removing MTD device #5 (rootfs_data) with use count 1 Jun 20 14:51:03 ed with return code: 265 Jun 20 14:51:10 I guess sysupgrade is failing? Jun 20 14:51:10 ubi foo? Jun 20 14:53:29 any way to find out why sysupgrade is aborting? Jun 20 14:57:40 third time's a charm... Jun 20 15:20:14 so ... lirc probably belongs in the packages feed. what about lirc kmod packages. should they go in the packages feed as well? Jun 20 15:21:13 * karlp is glad he doesn't hve to deal with lirc much, Jun 20 15:21:18 it awlays sounds like such a pig. Jun 20 15:27:09 karlp: I actually have never really used it before Jun 20 15:28:13 but it was pretty easy to get an IR emitter and receiver up and running with lirc_rpi Jun 20 15:29:32 figuring out what my AC remotes transmit however is probably going to be more difficult Jun 20 15:30:01 as what the buttons send via ir changes depending on what's on the display of the remote Jun 20 15:42:49 hm Jun 20 15:43:21 +632 bytes if I replace check_*() with a check_feature() function Jun 20 15:43:32 kinda horrible, but still 40 bytes Jun 20 15:45:25 and other things in the future following the same sort of process won't need explicit functions adding Jun 20 15:51:49 http://ix.io/1e54 Jun 20 15:51:56 pretty horrible :D Jun 20 16:40:59 mkresin: relocate-kernel uses a different loader implementation, the one from target/linux/generic/image/relocate Jun 20 16:43:01 Ideally, I'd like to see the 4 copies of lzma-loader and the relocate-loader for the different MIPS targets unified into a single source tree... but at the moment, all of them are separate Jun 20 16:48:33 neoraider: not sure if I ever stumbled over target/linux/generic/image/* Jun 20 16:49:23 looks like the ramips lzma-loader was some leftover till now Jun 20 17:23:17 hrm, usbip package comes from the kernel version, but needs patching for less than 4.14.50, and no patching for afterwards. Jun 20 17:23:21 that's going to be fun Jun 20 18:15:08 hello, i have question, i work at an isp, and we're planning to maybe deploy open-wrt on every routeur we sell. We would like to make few customisation and i'm not sure what is the easiest way to do it Jun 20 18:18:10 we would like to put permanent open port for remote management(both webgui and ssh) only for our ip address) and also maybe customize the web gui Jun 20 18:24:41 arist0v Better disable remote access by default and ensure to not expose it to the Internet if it's enabled (block this inbound port upstream) Jun 20 18:25:13 Most users do not expect their devices be accessed by anyone else. Jun 20 18:27:30 koops: we need to have acces to our client routeur and our client sign an agreement for this, so we could setup and avoid some wifi interferer Jun 20 18:28:05 I would never plug my devices into a box that can be remotely accessed by my ISP. Jun 20 18:28:23 for management you want opencpe or similar, netconf/yang type models Jun 20 18:28:34 or perhaps some of the openwisp stuff Jun 20 18:29:04 I would recommend requiring the user to approve any change manually, but that's tricky to implement. Jun 20 18:29:17 No unsupervised SSH access. Jun 20 18:29:22 koops: we already configure it manually on every router we sell, but we'Re looking for something more conveniant Jun 20 18:29:42 koops: and we limite it to traffic from our ip address Jun 20 18:29:43 you want a pull model Jun 20 18:29:57 pushing out config changes via ssh etc is so last year Jun 20 18:30:34 Yes, letting routers to pull latest settings is a better option. Jun 20 18:30:39 have the CPE either periodically poll for config changes, or use a persistent setup Jun 20 18:30:48 so you can signal changes Jun 20 18:30:53 Better than completely unsupervised access to webui and SSH. Jun 20 18:31:57 jwh: we need to be able to make change at any time, the ssh access is a more a back-up, but i think i found the way to do it myself so thank tou every one Jun 20 18:32:05 koops: unsupervised access is normal for CPE devices you get from your ISP Jun 20 18:32:33 Keep in mind that users run their LANs through these. Jun 20 18:33:00 At least most do, not everyone buys a $200 OpenWRT-compatible router like me. Jun 20 18:33:26 koops: i'm a bit security freak myself, so we just enable the minimum to prevent user to call us everytime we made change on our system(providing lot of client internet with 2.4ghz antenna) Jun 20 18:33:41 I would absolutely not enable ssh Jun 20 18:34:12 What about simply using 5Ghz? It doesn't penetrate walls and create lots of RF noise. Jun 20 18:34:13 unless you're terminating it in a vrf and the sshd is in a different network namespace Jun 20 18:34:26 One thick wall stops 5GHz completely. Jun 20 18:35:45 koops: we use technologe from 900 mhz to 5 ghz for client it depend of a lot thing Jun 20 18:39:44 What about simply setting Wi-Fi signal strength to minimum on each router? Jun 20 18:40:33 It should be enough for a small apartment. Jun 20 18:40:50 Like 10% of max strength should be fine. Jun 20 18:41:10 I don't remember any 900 MHz Wi-Fi standard. Jun 20 18:41:28 atheros used to make 11g cards on 900mhz Jun 20 18:41:42 I have one somewhere Jun 20 18:42:05 but uh, illegal to use here, used by mobile networks (so emergency services disruption etc) Jun 20 18:42:56 I had parents worrying about Wi-Fi effects on health and telling me to set antenna power to minimum, and it still covered the entire apartment. Jun 20 18:43:15 http://www.ciudadwireless.com/network_awpci900-_atheros_ar5414_900mhz--p-4299.html Jun 20 18:43:18 heh Jun 20 18:43:23 same chip I think Jun 20 18:43:41 jwh What exactly happens if you use 900 MHz? Jun 20 18:43:44 Arrested? Jun 20 18:43:53 probably prosecuted Jun 20 18:44:04 and you go to prison? Jun 20 18:44:11 especially when you interfere with an emergency call Jun 20 18:44:29 dunno, don't think anyone has ever been dumb enough to get caught Jun 20 18:44:30 heh Jun 20 18:47:49 it appears to be a civil offence (which doesn't preclude jail time) for a minor violtion Jun 20 18:47:52 violation Jun 20 18:48:01 but I expect it would be criminal if it resulted in loss of life Jun 20 18:50:28 Step 1. Hack into someone's router. Step 2. Make it create interference at emergency services frequencies. Step 3. Criminal charges for router's owner. Jun 20 18:50:31 Right? Jun 20 18:51:16 ignorance is no excuse, so absolutely Jun 20 18:51:28 One more reason to use secure, frequently updated router firmware like OpenWRT. Jun 20 18:51:55 but you can't generate arbitary interference with anything readily available Jun 20 18:51:58 so Jun 20 18:52:20 they're pretty tightly controlled by both the FCC and ETSI Jun 20 18:52:40 generally vendors want their devices to be approved by both, otherwise why bother Jun 20 18:53:26 So even if a router is hacked it can't send arbitrary radio signals? Jun 20 18:54:00 I doubt it is physically possible for your off the shelf netgear to transmit anything outside the 2ghz band its designed for Jun 20 18:54:16 not like it used to be when the atheros cards had wideband radios Jun 20 18:54:30 used to be able to do about 3.4ghz-5.9 Jun 20 18:54:51 for 11a ones Jun 20 18:55:07 got about 1.9-2.5 out of the 2ghz ones Jun 20 18:55:17 but that was maybe 12 years ago Jun 20 18:55:24 Would it be possible to make such card support 11ac with a firmware update? Jun 20 18:55:31 doubt it Jun 20 18:55:34 It seems to be able to transmit in 5GHz. Jun 20 18:56:16 if you could convince the firmware, and could change the filters Jun 20 18:56:21 etc Jun 20 18:57:11 Can a device compromised by extremely complex malware hack other devices over the air even if they are powered off? Jun 20 18:57:25 what Jun 20 18:57:31 Many people ask me this question, but I don't know myself. Jun 20 18:57:40 Mostly paranoid ones. Jun 20 18:57:44 how do you propose to interact with something without power Jun 20 18:58:07 generate electricity in circuits by radio waves somehow? Jun 20 18:58:11 calm down there tesla Jun 20 18:58:11 I don't know lol. Jun 20 18:58:44 https://en.wikipedia.org/wiki/Nikola_Tesla#Wireless_lighting Jun 20 18:58:45 :D Jun 20 19:01:58 So no? Jun 20 19:02:35 no Jun 20 19:02:55 theoretically under perfectly controlled conditions with specifically designed hardware, maybe Jun 20 19:03:20 I have a paranoid guy who sometimes needs something in his PC or router fixed. I installed OpenWRT for him. He is convinced that his PC and router and stuff are infected by badBIOS. Jun 20 19:03:35 probably Jun 20 19:03:44 how exactly are you going to provide enough power to the circuits by radio? Jun 20 19:03:50 field induction Jun 20 19:03:56 same way wireless charging works Jun 20 19:03:56 I have no idea, ask him. :-D Jun 20 19:03:58 you would need to be pretty close, no? Jun 20 19:04:01 yup Jun 20 19:04:14 you'd probably fry anything of value trying to do it at distance Jun 20 19:04:44 yeah Jun 20 19:05:03 not to mention you'd probably disrupt any radio communication anyway Jun 20 19:05:07 so pretty self defeating Jun 20 19:06:09 So, "hacking" by field induction is complete bullshit. I also thought it's bullshit, just wasn't able to explain why. Jun 20 19:06:31 its about as practical as badbios Jun 20 19:06:48 persistent malware? Jun 20 19:07:41 by means of ultrasonic radio comms Jun 20 19:07:43 was it actually ever confirmed or was the guy just tripping? Jun 20 19:07:48 via laptop speakers etc Jun 20 19:07:51 definitely tripping Jun 20 19:08:04 bit of fame Jun 20 19:08:10 Maybe both. It may exist but this guy is full of shit about having one on his PC. Jun 20 19:08:34 similar is theoretically doable, but not that specific claim Jun 20 19:09:02 That guy almost managed to persuade me by the way. Was almost too wary to get close to his PCs anymore. Jun 20 19:09:31 HOW CAN REGULAR PC SPEAKERS CREATE ULTRASOUND??? Jun 20 19:09:35 that guy needs to calm down and worry about more important stuff lol Jun 20 19:10:01 that's completely contrary to all knowledge of audiophile stuff I have. Jun 20 19:10:27 You need ridiculously expensive headphones or speakers to do that. Jun 20 19:11:16 For ultrasound or infrasound. Only the most expensive audiophile devices can play either of these. Jun 20 19:13:15 heh Jun 20 19:13:22 LOGIC PREVAILS Jun 20 19:14:37 do any of you guys have iptv? Jun 20 19:14:43 Yes. Jun 20 19:14:47 Yes Jun 20 19:16:05 is it a complex setup? Jun 20 19:17:15 my ISP used to pipe it through a separate vlan, but now it's both a saparate vlan and you have to set up some static routes because the decoder need to access the internet over the internet vlan Jun 20 19:17:37 Not really, I am using a dutch provider KPN. Which has a seperet IPTV vlan. Jun 20 19:17:48 Same setup. Jun 20 19:18:15 makes me wonder. why dont they just provide the needed internet access through the separate vlan? Jun 20 19:19:07 In case of KPN decoder, used for Netflix Jun 20 19:20:41 In my case I think it's because TV comes in through multicast on the TV vlan, but if you pause the program or watch something from the archive you need to access it through the internet. Jun 20 19:22:41 KPN decoder supports 2 modes. Directly connected to the IPTV vlan (Without internet apps support) or routed. Jun 20 19:22:41 Sennheiser HE 1 can play up to 100 000 Hz, but consumer non-audiophile headphones can't play ultrasound. Some can't even play the entire hearable range. Jun 20 19:23:00 aha Jun 20 19:23:21 Surprised he didn't claim it somehow communicated using a CPU cooler. Jun 20 19:24:34 Right now I use the ISP router/modem in bridge mode since it takes care of all the routing. I bridge the TV-port from the ISP router/modem over a separate SSID on my NETGEAR R7800 to my decoder. Jun 20 19:27:53 koops: fan modulation, now thats an idea Jun 20 19:28:23 if you can control it finely enough to generate the right pitch Jun 20 19:29:26 It would be interesting to get rid of the ISP router and just use the R7800 to do everything. However, if my ISP happens to change its setup while I'm not home a lot of people are going to be very angry with me, haha. Jun 20 19:31:17 heh Jun 20 19:33:13 If I will buy Netgear R7000 will it work with OpenWRT, just without Wi-Fi? Jun 20 19:34:11 I'm thinking about buying a Ubiquiti for WiFi, and not doing it on router. Jun 20 19:34:41 It seems most routers with powerful CPUs (for VPN) have Broadcom SoCs. Jun 20 19:34:58 Why the R7000? Jun 20 19:35:08 My budget is about $200. Jun 20 19:35:28 hmm Jun 20 19:35:31 I can invest more but really, I can build a PC for $300. Jun 20 19:35:56 And install pfSense on it, and it will be faster than any router on market. Jun 20 19:36:15 OpenWRT runs quite nicely on x86_64 too. Jun 20 19:36:28