**** BEGIN LOGGING AT Sun Mar 27 02:59:57 2016 Mar 27 10:49:17 freemangordon: v4.6-rc1-n900 branch is in linux-n900 tree Mar 27 10:54:16 Pali: great Mar 27 10:54:46 Pali: going to test it Mar 27 10:54:52 ok Mar 27 10:56:49 Pali: could you take a look at that gadget oops, I can't figure it out where it comes from Mar 27 10:57:31 Pali: I mean - why there is opps? Mar 27 10:57:35 *oops Mar 27 10:57:47 can you send me link to that email thread? Mar 27 10:57:52 looks like I was not CCed Mar 27 10:57:54 link? Mar 27 10:58:05 it is on linux-omap Mar 27 10:58:39 Pali: http://www.spinics.net/lists/linux-usb/msg135412.html Mar 27 11:00:07 "Unable to handle kernel paging request at virtual address 6c6c757a" Mar 27 11:00:08 that? Mar 27 11:03:07 hmmm.... 6c6c757a = "lluz" or "zull" Mar 27 11:07:03 freemangordon: looks like somebody freed pointer to common... Mar 27 11:07:18 from code around it looks like some race condition Mar 27 11:07:25 this is multithread code Mar 27 11:07:29 ah :-( Mar 27 12:05:52 Pali: yeah, but I was not able to figure out what happens Mar 27 12:06:17 who frees common and why Mar 27 12:06:42 Pali: fsg_unbind() is called twice for every lun Mar 27 12:07:00 also, I don;t see who terminates that thread Mar 27 12:07:51 that code is like spaghetti and being multithreaded doesn't help either :) Mar 27 12:09:10 Pali: hmm, wait, fsg_bind() is called 2 times as well, so the thread is created twice?!? Mar 27 12:33:30 Pali: seems like nokia_bind() is called twice, for some reason, so all gadgets are created 2 times Mar 27 16:21:33 is the yappari dude here? Mar 27 16:21:59 I'm worried about the security of my phone Mar 27 16:22:20 and the texts and shit, if people can spend enough time and get shit off my linux machine Mar 27 16:26:15 which they can and have Mar 27 16:30:34 how hard is whatsapp to hack Mar 27 17:13:23 fishbulb: what do you want to know? Mar 27 17:13:49 how hard is it to "hack" yappari Mar 27 17:13:55 from your point of view, realistically Mar 27 17:14:17 or whatsapp or whatever. I mean my end isn't encrypted Mar 27 17:14:34 whatsapp is harder, as conversations are ciphered point to point Mar 27 17:14:51 with a protocol (textsecure) which, as far as i know, hasn't been broken Mar 27 17:14:59 yappari doesn't implement this, however Mar 27 17:15:03 yeah I know Mar 27 17:15:13 so all security is based on SSL between your phone and WA servers Mar 27 17:15:27 so in that sense, it's as secure as any other ssl enabled website you visit Mar 27 17:15:46 it's a wrapper for a website? Mar 27 17:15:47 ceene: what about key exchange (in textsecure)? Mar 27 17:15:55 I didn't know that Mar 27 17:16:06 fishbulb: it's not a wrapper, but all communication is done via ssl Mar 27 17:16:12 or did you mean it's as secure.. ok Mar 27 17:16:19 you can't use a browser or anything similar to it to connect to whatsapp Mar 27 17:16:40 so from my end to the server is ssl Mar 27 17:16:53 bencoh: i don't know how it's done exactly Mar 27 17:17:02 and everyone else is encrypted, there have been less problems with dropped messages recently Mar 27 17:17:07 I noticed. Mar 27 17:18:03 TextSecure encryption protocol, which combines the Axolotl ratchet, prekeys, and a 3-DH handshake.[36] It uses Curve25519, AES-256, and HMAC-SHA256 as primitives Mar 27 17:18:15 aes256 Mar 27 17:18:20 that's pretty secure in theory Mar 27 17:18:30 if i'm not wrong, 3-DH is also what SSH uses to exchange keys Mar 27 17:19:12 weakest point on n900 is probably ssl support Mar 27 17:19:14 certificates Mar 27 17:19:15 and all that Mar 27 17:19:40 Pali: hmm, 4.6-rc1 seems pretty buggy :) Mar 27 17:19:53 did you try to boot it on the device? Mar 27 17:20:32 only in qemu Mar 27 17:22:00 Pali: http://pastebin.com/PYuw0SST Mar 27 17:22:28 "omapfb: dma_declare_coherent_memory failed" is also in qemu Mar 27 17:22:38 I had to disable et8ek8, otherwise it oopses Mar 27 17:22:57 ceene: the thing is, with ssh you either 1. have the remote fingerprint stored locally to check it, or 2. it's your first connection and you need to verify it manually (or blindly trust remote server) Mar 27 17:22:59 but why it fails? Mar 27 17:23:15 ceene: but I highly doubt whatsapp user are ever confronted to this :) Mar 27 17:23:39 which means that ... they can mitm the first exchange any way they want Mar 27 17:23:45 (they == whatsapp) Mar 27 17:24:07 i guess so, yes Mar 27 17:24:27 i mean, i'm always talking about a 3rd party intercepting your conversations Mar 27 17:24:38 i take it as a given that WA can read your messages if they want to Mar 27 17:24:52 not that i particularly believe they do, but that they do have the techinical ability is a given Mar 27 17:25:56 since they control the client (well, most of them), they can disable/modify/tamper with the encryption whenever they want, and there's nothing you can do about it Mar 27 17:26:35 that too Mar 27 17:28:32 Pali: seems like that one https://lists.01.org/pipermail/linux-nvdimm/2015-August/001796.html Mar 27 17:33:30 no, 3rd party Mar 27 17:33:39 I don't give a shit about whatsapp reading anything Mar 27 17:49:30 Pali: the bugger is 6b03ae0d42bfecf2536f7967d37e471d98c0b3d2 Mar 27 17:52:31 so need to remove DMA_MEMORY_MAP flag somehow? Mar 27 17:53:37 change it to DMA_MEMORY_IO Mar 27 17:54:03 It would've been really nice if someone explained what those flags are supposed to mean Mar 27 17:55:04 Pali: ^^^ Mar 27 17:55:48 with that change there is no more error in qemu, going to test on the device Mar 27 17:58:39 Pali: did you see my comments ^^^ that g_nokia probe() is called twice? Mar 27 17:58:50 yes Mar 27 17:58:58 but do not know why.... Mar 27 18:00:24 Pali: ok, the DMA_MEMORY_IO change works on the device as well, right now I have "9" playing in mediaplayer :) Mar 27 18:00:56 which means tidspbridge works as well, to the extend it did with 4.5 ofc Mar 27 18:10:29 \o/ Mar 27 18:10:54 who the man Mar 27 18:10:57 freemangordon is Mar 27 18:11:16 and Pali Mar 27 18:12:30 we are very lucky to have both of them working on n900 instead of something that'll make them rich and famous :) **** ENDING LOGGING AT Mon Mar 28 02:59:57 2016