**** BEGIN LOGGING AT Mon Mar 04 02:59:57 2019 Mar 04 08:54:41 KanjiMonster: thank you! I have updated the documentation. Mar 04 10:04:15 pkgadd: Thanks for reporting. Rene__ also reported some errors, and I could simulate it. I'll probably bump 4.9 and 4.14. but 4.19 will have to wait until it's spotless again :) Mar 04 12:13:49 Hi all, I'm trying to understand how mounting overlay works during bootup. It mostly makes sense, init executes /lib/preinit.sh and one of the hooks runs mount_root. mount_root mounts the rw volume, then mounts overlayfs with said volume as upper directory then performs a pivot_root to overlayfs. Mar 04 12:14:15 When preinit completes procd is executed inline to override init as the pid1 process. Mar 04 12:15:31 The bit that confuses me is that you don't cd to rw volume before pivot_root not do you chroot to it afterwards. The man page for pivot_root suggests you should do that to guarenteee that cwd and / change for the calling process. But OpenWRT doesn't seem to do that so how does procd have the correct cwd and /? Mar 04 15:46:49 don't forget to give feedback on this ridiculous EU idea to lock down radio equipment Mar 04 15:47:09 https://blog.mehl.mx/2019/protect-freedom-on-radio-devices-raise-your-voice-today/ Mar 04 19:59:17 stintel: should we give feedback as OpenWrt? Mar 04 20:08:54 Hauke: we have time till midnight. go for it, if you can! Mar 04 20:10:58 ok Mar 04 20:11:18 if someone else also wants we can have two or more ;-) Mar 04 20:49:01 Hauke: I did personally Mar 04 20:49:21 but maybe as OpenWrt would be a good idea Mar 04 20:59:49 stintel: lynxis http://piratepad.net/Bm46BUjFvj Mar 04 21:17:35 are you drafting a formal remark for the EU proposal? Mar 04 22:03:33 another thing, being closed source won't prevent from violating regulations, people can still unpack the firmware, change the regulatory domain, repack the image and flash their device Mar 04 22:04:05 same as it is possible to add/update stuff in a vendor-provided image Mar 04 22:04:36 Is /etc/os-release or /etc/openwrt_release preferred for build-related info? Mar 04 22:05:16 (for _reading_ build-related info) Mar 04 22:06:16 Mister_X: you can't do that with signed images Mar 04 22:06:30 true Mar 04 22:06:47 do many vendor provide signed images? Mar 04 22:07:22 they would love to Mar 04 22:07:52 what prevents them from doing it? Mar 04 22:08:16 Mister_X: They're too stupid Mar 04 22:08:49 It didn't come in the Qualcomm / MTK toolkit Mar 04 22:08:51 cost issue Mar 04 22:09:23 isn't that regulation becoming irrelevant with all the firmware blobs in 802.11n/ac/ax? Mar 04 22:09:37 incorporating those limitations Mar 04 22:09:43 No benefit to them for consumer routers. In contrast, for cell phones, locked bootloaders reduce support costs Mar 04 22:10:16 Did I miss something -- the EU is proposing a roadmap, not any regulations at this point, at least from what I read. Mar 04 22:10:48 Yes, commenting on what they should consider is valuable, but I didn't see any proposed regulation at all Mar 04 22:12:09 If I were an EU subject, I'd be making an argument based on the fundamental right to privacy, as enshrined in the GDPR, for example, and the manufacturer's failure to promptly update firmware Mar 04 22:13:17 but they're pretty much against that right to privacy AFAIK Mar 04 22:13:33 or, for that matter, to provide firmware in the first place that can robustly protect the privacy of the user. Not talking VPN here, just security of the system Mar 04 22:13:38 but the second part of your argument is great Mar 04 22:15:10 openwrt also extend product life, which reduces electronic waste Mar 04 22:16:00 Chapter 1, Article 1 of the GDPR, "This Regulation protects fundamental rights and freedoms of natural persons and in particular their right to the protection of personal data. " Mar 04 22:16:39 That's pretty clear precedent Mar 04 22:16:43 yep Mar 04 22:17:06 but if you're mentioning privacy, you need to explain that it is regarding that point Mar 04 22:17:26 "right to protection of personal data" Mar 04 22:17:28 because often times, people think that it relates to VPN and people hiding their stuff Mar 04 22:18:07 If you're requiring data processors to use best practices in protecting others' data, I should be able to protect my own data using best practices Mar 04 22:18:40 (without even talking about VPNs or encryption -- just penetration) Mar 04 22:18:43 another thing that openwrt does is often extend features; vendors limit to a small set of features, openwrt is more flexible (but I'm not sure if it can be used in any way) Mar 04 22:19:34 Also, I linked the FCC consent degree against TP-Link, which explicitly discusses third-party firmware Mar 04 22:20:33 Using their arguments likely will have more weigh than a "hacker" saying they want to run their own firmware because its "better" (put yourself in the shoes of the commission) Mar 04 22:22:37 https://www.fcc.gov/document/fcc-reaches-200000-settlement-tp-link-usa-corporation Mar 04 22:24:20 As part of the Compliance Plan, TP-Link also agrees to take the steps set forth herein to support the development of software by third-parties, including open-source software, for use with its routers, which comply with the U-NII security requirements. Mar 04 22:24:51 jeffsf, maybe want to reach out to Glinet Mar 04 22:25:13 their firmware is based on openwrt Mar 04 22:25:35 they may have some additional comments/ideas Mar 04 22:26:53 Certainly worth input, but I haven't seen anything that says that the EU is planning on prohibiting third-party firmware, just that they are planning on examining what regulations might be appropriate to propose Mar 04 22:28:26 Read the scope, "Reconfigurable Radio Systems" (aka software-defined radios) Mar 04 22:28:51 “This is made possible by the software implementation of electronic components that were typically hardware (e.g. diodes, switches, mixers, filters, demodulators).” Mar 04 22:29:39 I think they do not know that they are banning thrid party firmware Mar 04 22:30:11 True, making sure that it doesn't impact the market broadly is important Mar 04 22:31:23 if I remember correctly, some cellphone equipment use SDR Mar 04 22:31:38 “In recital (19) of the same Directive, the co-legislators also stressed that Verification by radio equipment of the compliance of its combination with software should not be abused in order to prevent its use with software provided by independent parties. “ Mar 04 22:31:41 professional cellphone equipment Mar 04 22:31:42 " Mar 04 22:32:40 Which is the same position the FCC took against TP-Link Mar 04 22:33:58 jeffsf: they can sign our software if we pay them 20K$ per software, they need 2 weeks to test this so it is reasoable Mar 04 22:35:29 and we would be liable if something is wrong Mar 04 22:43:13 so, every release would cost $20K? Mar 04 22:43:46 per device Mar 04 22:44:17 this is just an example, how one can comply with this sentence Mar 04 22:45:39 what does that test involve? Mar 04 22:46:45 I'm wondering if would be possible to make something to drive down the cost drastically Mar 04 22:47:23 it is published: https://ec.europa.eu/info/law/better-regulation/initiatives/ares-2018-6621038/feedback/F240894_en?p_id=380919 Mar 04 22:47:51 Mister_X: many vendors are also not intrested to run something else on their hardware Mar 04 22:48:43 Mister_X: currently we try to exclude home routers from these requirements Mar 04 23:01:29 anyone know how to figure out a map of the partitions on a device? I'm looking at /proc/mtd but that only gives me sizes, not offsets Mar 04 23:04:25 ah dmesg Mar 04 23:04:26 nvm Mar 04 23:06:51 mangix: personally I find the dmesg output to be the most readable form, but that's not always available (ring buffer for log messages too small) Mar 04 23:09:16 pkgadd: i'm working on porting a device from ar71xx to ath79. turns out there's a similar device already ported. Mar 04 23:12:32 even better, I have a device waiting here to be installed (meaning I can't open- or break it) running the vendor software and it's a tad confusing to read out the actual sizes of the installed NOR and NAND flash (NOR seems to be 32 MB, but NAND is harder to answer, with /proc/mtd not really distinguishing between them); and the dmesg buffer being too small for catching those values Mar 04 23:47:12 tada, Linux 5.0 Mar 05 00:04:20 3.41 you mean? :) Mar 05 00:11:40 from that stance you could call it 2.6.82 :-P Mar 05 01:31:14 Hauke: thank you for taking care! Mar 05 02:31:10 oh this is hillarious. I can't get the ethernet port to work, only to find that qca953x.dtsi is forcing the link to 1G Mar 05 02:59:39 gch981213: ping **** ENDING LOGGING AT Tue Mar 05 02:59:57 2019