*** NIN101 has quit IRC | 00:04 | |
FatPhil | Ah, that looks like Pali's tree, and freemangordon's is a fork from it | 00:09 |
---|---|---|
*** nox- has joined #maemo-ssu | 00:33 | |
*** Martix has joined #maemo-ssu | 00:44 | |
*** amiconn has quit IRC | 01:48 | |
*** amiconn has joined #maemo-ssu | 01:49 | |
*** discopig has quit IRC | 02:21 | |
*** LaoLang_cool has joined #maemo-ssu | 02:30 | |
*** Martix has quit IRC | 02:35 | |
*** nox- has quit IRC | 03:40 | |
*** dos1 has joined #maemo-ssu | 03:43 | |
*** jonwil has joined #maemo-ssu | 03:46 | |
*** kolp_ has joined #maemo-ssu | 03:57 | |
*** kolp has quit IRC | 04:01 | |
*** kolp_ is now known as kolp | 04:02 | |
*** kolp has quit IRC | 04:07 | |
*** kolp has joined #maemo-ssu | 04:07 | |
*** dos1 has quit IRC | 04:09 | |
*** kolp_ has joined #maemo-ssu | 04:25 | |
*** wmarone has quit IRC | 04:27 | |
*** kolp has quit IRC | 04:30 | |
*** wmarone has joined #maemo-ssu | 04:30 | |
*** kolp_ is now known as kolp | 04:36 | |
*** LauRoman has joined #maemo-ssu | 04:41 | |
*** LauRoman has quit IRC | 05:14 | |
*** Milhouse has quit IRC | 05:22 | |
*** kolp has quit IRC | 05:24 | |
*** amiconn has quit IRC | 05:41 | |
*** amiconn_ has joined #maemo-ssu | 05:41 | |
*** amiconn_ is now known as amiconn | 05:41 | |
*** ototo has quit IRC | 05:44 | |
*** g3kk3r has quit IRC | 05:44 | |
*** tg has quit IRC | 05:44 | |
*** DrCode has quit IRC | 05:44 | |
*** amizraa has quit IRC | 05:44 | |
*** g3kk3r has joined #maemo-ssu | 05:44 | |
*** ototo has joined #maemo-ssu | 05:44 | |
*** ototo has quit IRC | 05:44 | |
*** ototo has joined #maemo-ssu | 05:44 | |
*** tg has joined #maemo-ssu | 05:45 | |
*** tg has quit IRC | 05:45 | |
*** tg has joined #maemo-ssu | 05:45 | |
*** DrCode has joined #maemo-ssu | 05:55 | |
*** amizraa has joined #maemo-ssu | 05:55 | |
*** DrCode has quit IRC | 05:55 | |
*** DrCode has joined #maemo-ssu | 05:56 | |
*** discopig has joined #maemo-ssu | 06:00 | |
*** aap has quit IRC | 06:38 | |
*** Milhouse has joined #maemo-ssu | 07:02 | |
*** aap has joined #maemo-ssu | 07:11 | |
*** Vlad_on_the_road has joined #maemo-ssu | 07:32 | |
*** Vlad_on_the_road has quit IRC | 07:41 | |
*** dafox has joined #maemo-ssu | 07:52 | |
*** Vlad_on_the_road has joined #maemo-ssu | 08:08 | |
*** Vlad_on_the_road has quit IRC | 08:24 | |
*** dafox has quit IRC | 08:38 | |
*** Pali has joined #maemo-ssu | 08:48 | |
*** luf has joined #maemo-ssu | 09:01 | |
*** Pali has quit IRC | 09:18 | |
freemangordon | FatPhil: if you manage to boot fremantle with 3.12 on you devel device and wonder what to start doing with it, you may want to look why after a reboot from 2.6.28 to 3.12 a couple of devices connected to i2c fail to start :) | 09:52 |
*** Pali has joined #maemo-ssu | 09:54 | |
freemangordon | FatPhil: and yes, my tree is a clone of Pali's, but Pali's lag a little because of his faulty laptop | 09:54 |
FatPhil | freemangordon: good to know. I cloned yours, with linux as a --reference, I'll try to pull daily. How often to you merge? | 09:56 |
freemangordon | it is pali who pulls upstream, iiuc on a release | 09:56 |
Pali | no merge | 09:56 |
Pali | I only rebasing patches on rc/release versions | 09:57 |
Pali | so I always see patches which are not in upstream... | 09:57 |
freemangordon | Pali: are there lots of them? | 09:58 |
Pali | yes | 09:58 |
freemangordon | (excluding drivers) | 09:58 |
Pali | I think all are drivers or board code | 09:58 |
freemangordon | well, not that bad then | 09:59 |
Pali | only small of them patching something other | 09:59 |
freemangordon | good | 09:59 |
freemangordon | sailus: ping | 10:00 |
*** Martix has joined #maemo-ssu | 10:03 | |
Pali | freemangordon: did you already looked at DT? | 10:10 |
freemangordon | Pali: no, I am waiting for something to come from upstream | 10:11 |
freemangordon | in 3.13 there should be | 10:11 |
Pali | ok | 10:11 |
Pali | then we will wait for 3.13 | 10:11 |
freemangordon | I'll send a couple of mails about broken omapfb later on | 10:12 |
Pali | so now only need to fix snd (with jack detection), omapfb and camera, right? | 10:12 |
freemangordon | yep | 10:13 |
freemangordon | and gpios | 10:13 |
Pali | good | 10:13 |
Pali | and what about power management? | 10:13 |
freemangordon | ooh, and LEDs | 10:13 |
Pali | ah leds, I will look at them | 10:13 |
freemangordon | Pali: I think we should have everything working first | 10:13 |
freemangordon | touchscreen is not disabled, so it is hard to say what is the state of PM | 10:14 |
freemangordon | hmm, maybe I should look at that issu | 10:14 |
freemangordon | *issue | 10:14 |
Pali | touchscreen is not disabled when lock screen? | 10:14 |
freemangordon | yes | 10:14 |
Pali | maybe problem with slide gpio and mce | 10:14 |
Pali | or maybe missing some touchscreen sysfs entry | 10:15 |
freemangordon | no, display is powered off | 10:15 |
freemangordon | but TS still active | 10:15 |
freemangordon | :nod: (missing sysfs entry) | 10:15 |
freemangordon | i'll look at it | 10:16 |
FatPhil | Pali: if you rebase, then you're not fast-forwardable. You're not a stable tree upon which to base anything else. | 10:18 |
FatPhil | http://changelog.complete.org/archives/586-rebase-considered-harmful | 10:18 |
FatPhil | See the first bit in bold | 10:19 |
FatPhil | It's great for solo work, but not when you're sharing your tree | 10:19 |
Pali | FatPhil: for each kernel version I have separate patch set | 10:20 |
Pali | and rather I will use quilt for n900 patches, but it is not easy for use | 10:20 |
FatPhil | quilt was invented when git didn't exist. there's no logical reason to use it now git exists. | 10:21 |
Pali | for each kernel version I have separe branch | 10:21 |
FatPhil | OK, at least that helps keep things sane | 10:21 |
Pali | I need to see all patches which are not in upstream | 10:21 |
*** Martix has quit IRC | 10:22 | |
Pali | and I also need to generate formated patches for sending to upsteam | 10:22 |
Pali | which means I must rebase patches on top of some new kernel version | 10:22 |
Pali | fast forward and merge will totally broke it | 10:23 |
*** Martix has joined #maemo-ssu | 10:23 | |
Pali | and then there will be big mess about patches... | 10:23 |
Pali | when I rebasing patches on top of new kernel version, I doing some patch simplification or merged more patches to one if that make sense | 10:24 |
Pali | so I have clean patches... | 10:24 |
Pali | and I never doing push --force, so repository is never broken | 10:24 |
Pali | you always can do pull fast forward | 10:24 |
Pali | but sometimes you need to switch to new branch | 10:25 |
freemangordon | hmm: "/sys/devices/platform/omap2_mcspi.1/spi1.0/disable_ts" | 10:25 |
FatPhil | whose touchscreen driver are you running? | 10:27 |
freemangordon | upstream | 10:27 |
FatPhil | does it have a threaded irq? | 10:27 |
freemangordon | NFC | 10:27 |
FatPhil | OK, aaro's not mine | 10:27 |
freemangordon | but it should have, as it is forced | 10:27 |
freemangordon | (threaded irq) | 10:27 |
FatPhil | there were no threaded irqs in .28, when I was maintaining tsc2005.c | 10:28 |
*** Martix has quit IRC | 10:28 | |
freemangordon | in 3.12 all are threaded, unless you specify a flag in request_irq, iiuc | 10:28 |
FatPhil | Dmitry Torokhov's done some big cleanups to that code too, it's changed quite a lot | 10:29 |
jonwil | btw did we end up solving the problem of that one driver where the kernel guys refused to accept changes because "its Nokia's Driver and only Nokia can change it"? | 10:30 |
freemangordon | jonwil: there is no problem, Pali wanted to simplify the things | 10:31 |
jonwil | ok | 10:31 |
freemangordon | iirc | 10:31 |
Pali | jonwil: it is still a problem | 10:31 |
jonwil | ok | 10:31 |
Pali | in usb subsystem | 10:31 |
FatPhil | jonwil: that was a patch from Pali that Felipe rejected as it created potential breakage for legacy userspace | 10:32 |
FatPhil | IIRC | 10:32 |
jonwil | ok | 10:32 |
Pali | there was no userspace breakage | 10:32 |
freemangordon | :nod: | 10:32 |
Pali | but I will send that patch for comments again | 10:32 |
Pali | on top of 3.12-rc1 | 10:33 |
FatPhil | Didn't it change the userspace API? | 10:33 |
Pali | in 3.10 was g_nokia totally broken | 10:33 |
Pali | no, it did not changed any api | 10:33 |
Pali | it added new optional support | 10:33 |
Pali | which can be at runtime enabled/disabled | 10:33 |
Pali | freemangordon: is g_nokia working in 3.12? | 10:34 |
Pali | is pc suite mode working? | 10:34 |
Pali | is mass storage working? | 10:34 |
freemangordon | g_nokia is broken, at least in bootmenu | 10:34 |
Pali | can you test it? | 10:34 |
freemangordon | ok, I will | 10:34 |
Pali | ok | 10:34 |
freemangordon | Pali: would bq start charging if device booted with charger attached? | 10:36 |
freemangordon | bq driver that is | 10:36 |
FatPhil | ah, this thread... http://lkml.org/lkml/2013/1/22/375 | 10:37 |
Pali | freemangordon: should | 10:38 |
*** luf has quit IRC | 10:39 | |
freemangordon | well, well, the path for TS is different | 10:44 |
freemangordon | BUT - there is no "disable" sysfs entry iiuc | 10:44 |
Pali | freemangordon: so only sysfs entry is missing? | 10:44 |
freemangordon | Pali: and path differs too | 10:45 |
freemangordon | "/sys/devices/platform/omap2_mcspi.1/spi_master/spi1/spi1.0" is in 3.12 | 10:45 |
Pali | you should use /sys/class | 10:45 |
freemangordon | "/sys/devices/platform/omap2_mcspi.1/spi1.0/" is what mce looks for | 10:45 |
freemangordon | Pali: me? :P | 10:46 |
Pali | can you check if you have it in /sys/class/ too? | 10:46 |
freemangordon | "/sys/class/spi_master/spi1" | 10:46 |
freemangordon | that is linked to "/sys/devices/platform/omap2_mcspi.1/spi_master/spi1" | 10:47 |
Pali | some other class name (like input/touchscreen/....)? | 10:48 |
FatPhil | commit 5cb81d19bae47adcb073a5e5a3bc40dd252f239e | 10:49 |
FatPhil | hopefully revert to solve "so only sysfs entry is missing?" | 10:49 |
FatPhil | pure guess | 10:49 |
Pali | on kp52 I have /sys/class/spi_master/spi1/device/spi1.0 | 10:50 |
freemangordon | not in 3.12 | 10:51 |
Pali | and what is on 3.12? | 10:52 |
freemangordon | Pali: ^^^ | 10:52 |
freemangordon | "/sys/class/spi_master/spi1" | 10:53 |
freemangordon | linked to "/sys/devices/platform/omap2_mcspi.1/spi_master/spi1" | 10:53 |
Pali | kp52: /sys/class/spi_master/spi1/device/spi1.0/disable_ts | 10:53 |
Pali | in 3.12 there is? /sys/class/spi_master/spi1/disable_ts | 10:54 |
freemangordon | no, see the commit posted by FatPhil | 10:54 |
Pali | link? | 10:54 |
freemangordon | "Input: tsc2005 - remove 'disable' sysfs attribute" | 10:54 |
Pali | I do not have git tree on n900 :D | 10:55 |
freemangordon | this is the title ^^^ | 10:55 |
freemangordon | lets see if we can use some "power" sysfs entry | 10:55 |
FatPhil | http://www.spinics.net/lists/linux-input/msg14575.html | 10:55 |
FatPhil | Pali: ^^^ = that patch | 10:55 |
Pali | ok | 10:55 |
FatPhil | Who was interested in the AES h/w acceleration | 10:56 |
Pali | if we can offload aes from cpu to other dedicated hw, then it can be really usefull | 10:58 |
Pali | e.g. for luks | 10:58 |
freemangordon | Pali: FatPhil: any clue how to power down(disable) a device using sysfs entry? | 10:59 |
FatPhil | Pali: http://privatepaste.com/97443f891d | 10:59 |
Pali | write something to power/level | 10:59 |
Pali | and if level does not exits, then there is no way | 11:00 |
FatPhil | freemangordon: I liked the old sysfs way, but I have a 2.6.28 brain | 11:00 |
freemangordon | Pali: what needs to be done so "level" to appear? | 11:00 |
freemangordon | FatPhil: same here :D | 11:00 |
FatPhil | I'm not sure why Dmitry removed the functionality from tsc2005 before adding a generic replacement | 11:00 |
Pali | FatPhil: about that secure code: it is not possible to enable it from non secure side via smc calls? | 11:01 |
Pali | like IBE bit for thumb errata | 11:02 |
Pali | freemangordon: report bug to maintainer for reverting old sysfs entry for compatibility mode (or to implemented new support) | 11:03 |
FatPhil | Maybe someone could ask him why he broke userspace ;-) | 11:03 |
Pali | right | 11:03 |
*** freemangordon_ has joined #maemo-ssu | 11:06 | |
*** freemangordon has quit IRC | 11:06 | |
FatPhil | A very cheesy way of "disabling" the ts is to just close the device node, of course | 11:07 |
FatPhil | who's the client that relies on ts_disable? | 11:07 |
freemangordon_ | FatPhil: mce :D | 11:08 |
freemangordon_ | FatPhil: (power outage here) is there a way to open that firewall by using SMC? | 11:09 |
*** freemangordon_ is now known as freemangordon | 11:09 | |
freemangordon | FatPhil: also, what exactly do you mean by saying "firewall"? all I know is some co-processor registers | 11:10 |
Pali | omap firewall from that pastebin | 11:11 |
freemangordon | what IS "omap firewall", I've never seens such a term in ARM docs | 11:11 |
FatPhil | freemangordon: have you even seen the secure mode docs ;-) | 11:12 |
freemangordon | no, but co-pro regs are well documented in public docs | 11:13 |
FatPhil | My understanding, and I've never ever looked at secure mode, is that the firewall is just a name for rejecting certain requests. If it's not configured in secure mode to enable those calls, you can't call them. | 11:14 |
freemangordon | which is exactly what some regs in system control coprocessor do | 11:15 |
freemangordon | though mshield is missing from the public docs afaik | 11:16 |
jonwil | so much omap bits not made public and only revealed under NDA | 11:17 |
FatPhil | 90% of the kernel developers didn't have access to that bit of the TI docs | 11:19 |
FatPhil | Even the TI onsite support dude didn't have access to the TrustZone docs, it was that secretive | 11:19 |
*** dhbiker has quit IRC | 11:19 | |
Pali | there is some docs about arm trustzone | 11:20 |
Pali | on arm website | 11:20 |
Pali | only need to register | 11:20 |
jonwil | Probably not the secret bits | 11:21 |
freemangordon | HW AES shouldn;t be that secret | 11:21 |
jonwil | That Jazelle stuff for Java speedup is also totally secret :( | 11:22 |
freemangordon | yeah :( | 11:22 |
jonwil | I think the Java thing was a deal with Sun to make sure that only Sun approved JVMs could use the hardware acceleration | 11:22 |
FatPhil | now deprecated, thank ${DEITY} | 11:24 |
Pali | there is thumb-ee ISA for jit compilers | 11:24 |
jonwil | Surprised no-one didn't just reverse engineer Jazzelle | 11:26 |
jonwil | Wouldn't be the first time someone reverse engineered undocumented instruction sets :P | 11:26 |
Pali | because of java? | 11:27 |
Pali | why somebody want to have java on mobile??? | 11:27 |
freemangordon | jonwil: nothing to RE afaik. jazelle is something like FP on 8088 iiuc | 11:27 |
freemangordon | you either have HW support, or you get an interrup/fault/whatever | 11:27 |
*** dhbiker has joined #maemo-ssu | 11:27 | |
Pali | rather to have native code as some hw support for java jvm | 11:28 |
freemangordon | jonwil: iiuc, you have some java bytecode instructions supported in HW | 11:28 |
*** Pali has quit IRC | 11:30 | |
jonwil | Jazzelle came about back when J2ME was king and the CPUs in phones were less powerful than they are today | 11:32 |
jonwil | so it made sense to support it | 11:32 |
jonwil | http://en.wikipedia.org/wiki/Jazelle indicates that using it isn't as simple as it sounds | 11:33 |
freemangordon | hmm, no "level" here https://gitorious.org/linux-n900/freemangordons-linux-n900/source/777f4ace4c58799575391833db48050ceefe6fd6:Documentation/ABI/testing/sysfs-devices-power | 11:34 |
*** kolp has joined #maemo-ssu | 11:45 | |
FatPhil | freemangordon: Doh - I'm being thick, the very next patch (backwards in time) shows what the correct power interface is - open/close! commit 0b950d3d7ce4c1e870b8efc4ae0faaf0ef53532c | 11:58 |
freemangordon | FatPhil: yep, saw that | 11:59 |
*** joshgillies has joined #maemo-ssu | 11:59 | |
freemangordon | but I don;t know who opens it | 11:59 |
FatPhil | X? | 12:00 |
freemangordon | could be | 12:00 |
freemangordon | but doesn't help much :D | 12:00 |
FatPhil | lsof should say | 12:00 |
freemangordon | Nokia-N900:/dev/input# lsof ts | 12:02 |
freemangordon | COMMAND PID USER FD TYPE DEVICE SIZE NODE NAME | 12:02 |
freemangordon | Xorg 1581 root 11u CHR 13,67 6903 event3 | 12:02 |
freemangordon | hald-addo 1664 root 6r CHR 13,67 6903 event3 | 12:02 |
freemangordon | mce 1948 root 10r CHR 13,67 6903 event3 | 12:02 |
freemangordon | open/close won't help | 12:02 |
FatPhil | yup | 12:02 |
FatPhil | Dmitry needs mailing | 12:02 |
freemangordon | could you do it? | 12:02 |
FatPhil | sure thing | 12:03 |
freemangordon | copy me and Pali please | 12:03 |
FatPhil | That's a possibility. As long as it doesn't immediately get dismissed as just being those troublemakers again... | 12:06 |
freemangordon | FatPhil: hmm? | 12:07 |
freemangordon | who are the troublemakers? | 12:07 |
freemangordon | :D | 12:07 |
freemangordon | ooh, those die hard n900 guys who want the kernel to work on a real HW. I see... | 12:08 |
freemangordon | FatPhil: BTW I don;t see what else could be used | 12:09 |
FatPhil | Well, people who swear at Felipe Balbi might be considered troublemakers, for example | 12:09 |
freemangordon | don;t copy Pali then ;) | 12:09 |
freemangordon | FatPhil: any idea what else could be used instead of disable sysfs entry? | 12:10 |
FatPhil | I think revert the patch is the best option | 12:11 |
FatPhil | It's lunchtime here though. I'll write the mail with a full belly | 12:12 |
freemangordon | sounds sane | 12:12 |
FatPhil | I got as far as: | 12:18 |
FatPhil | Back in http://www.spinics.net/lists/linux-input/msg14575.html [PATCH 17/17] Input: tsc2005 - remove 'disable' sysfs attribute | 12:18 |
FatPhil | you broke Nokia n900 userspace. | 12:18 |
FatPhil | But decided that was not the polite approach | 12:18 |
freemangordon | yeah, you'd better have something for lunch :) | 12:18 |
FatPhil | need food! but before I get food, I need socks!!! (moved back to tallinn last weekend, everything's chaos here) | 12:21 |
*** luf has joined #maemo-ssu | 12:31 | |
*** Pali has joined #maemo-ssu | 12:35 | |
*** lizardo has joined #maemo-ssu | 13:35 | |
*** DrCode has quit IRC | 14:28 | |
*** DrCode has joined #maemo-ssu | 14:31 | |
*** xmlich02 has joined #maemo-ssu | 14:34 | |
*** joshgillies has quit IRC | 14:41 | |
*** joshgillies has joined #maemo-ssu | 14:45 | |
Pali | FatPhil, freemangordon: when you sent those mails to lkml (without cc), please at least send links here | 14:56 |
FatPhil | Pali: will do | 15:03 |
FatPhil | HAs anyone ever seen the UI "shrunk"? | 15:04 |
FatPhil | Fulol functional, just only taking up 2/3rds of the screen | 15:04 |
FatPhil | s/fulol/fully | 15:04 |
merlin1991 | FatPhil: yep | 15:04 |
merlin1991 | reboot and it's gone | 15:04 |
merlin1991 | we don't know the exact cause | 15:05 |
merlin1991 | btw to use it, the touchscreen still is used as if the ui were "fullscreen" :D | 15:05 |
FatPhil | yup, worked that out! | 15:21 |
FatPhil | The weird thing is that the swipe-to-unlock is the right size! | 15:21 |
kolp | FatPhil: For me it happend when I locked the device right after the wallpaper showed at boot time. When I then unlock a little later the scrren is 2/3rds like you see | 15:31 |
Pali | freemangordon: is there public documentation about omap3 smc calls? | 15:39 |
Pali | to check if omap hw rand driver is ok? | 15:39 |
*** DrCode has quit IRC | 15:51 | |
*** DrCode has joined #maemo-ssu | 15:53 | |
*** xmlich02 has quit IRC | 16:58 | |
*** Martix has joined #maemo-ssu | 17:05 | |
FatPhil | oi, silly bot, what was the URL I told you to pass on to Pali... | 17:19 |
jonwil | wouldn't that be in the chanlog? :P | 17:19 |
FatPhil | Yeah, but I want the bot to feel useless, lest it gets a sense of superiority to humans | 17:21 |
Pali | irc memoserv is for offline messages .-) | 17:22 |
FatPhil | YRC, many thanks! | 17:26 |
FatPhil | jonwil: have you decompiled nolo? | 17:29 |
jonwil | nope, I wouldn't know where to begin as its a binary blob rather than an ELF executable | 17:30 |
jonwil | and obviously I cant decompile it since its ARM and not x86 | 17:30 |
Pali | you can embed binary blob into elf :-) | 17:30 |
jonwil | but no I haven't disassembled it | 17:30 |
jonwil | why the interest in nolo? | 17:30 |
Pali | just compile empty file to elf object file | 17:31 |
Pali | and then objcopy nolo into that elf object | 17:31 |
Pali | and then objdump -D for disassembling working fine | 17:31 |
jonwil | its not about loading the binary into IDA its about knowing where in the address space the different parts of the blob need to sit, its about knowing whats instructions and whats data etc | 17:32 |
*** lizardo has quit IRC | 17:32 | |
jonwil | also I have no reason to care much about nolo | 17:32 |
jonwil | Its not going to be usable on Neo900 in any case | 17:32 |
jonwil | since nolo is tied to the specific OMAP SoC in N900 | 17:32 |
jonwil | so again I say, why the interest in NoLo? | 17:39 |
freemangordon | jonwil: I suspect because of AES | 17:43 |
*** lizardo has joined #maemo-ssu | 17:44 | |
freemangordon | hmm, but why nolo? | 17:44 |
freemangordon | Pali: there is something mentioned in public 3430 TRM, lemme check what it is | 17:45 |
Pali | freemangordon: ok | 17:45 |
Pali | some doc for omap3 rom rng can be usefull | 17:45 |
jonwil | Personally I would be less inclined to trust hardware cryptography given all the recent NSA stuff (and given that Texas Instruments is a big USA company that would probably bend over backwards if the NSA wanted them to add backdoors to their parts) | 17:46 |
jonwil | and I know a few security researchers and crypto geeks who would agree with the idea of being suspicious of hardware crypto and RNGs and stuff in mass-market hardware | 17:47 |
Pali | seed from hw rng for /dev/urandom should be safe | 17:48 |
jonwil | at least with a software implementation in, say, openssl, its possible to audit that implementation | 17:48 |
Pali | and good | 17:48 |
freemangordon | Pali: all is about L2 and AUX CTL, | 17:48 |
freemangordon | SWPU223M, p.3382 | 17:49 |
Pali | ok | 17:49 |
Pali | I did not find anything on internet about omap3 rng (expect info in maemo kernel) | 17:50 |
Pali | anybody has doc about smc/secure part? | 17:51 |
jonwil | TI seems to want to keep all that stuff secret | 17:55 |
jonwil | including security features | 17:55 |
jonwil | hardware RNG/crypto | 17:55 |
jonwil | etc etc | 17:55 |
freemangordon | jonwil: most probably they are missing in GP devices | 17:55 |
jonwil | I doubt anyone who has that doc will be willing to share public (it likely has a lot of "confidential TI information" warnings at the top of irt) | 17:56 |
Pali | yes, gp devices does not have them | 17:56 |
Pali | but why is api big secret? if you want to use it - it must be in kernel and that is open source... | 17:57 |
freemangordon | Pali: how is HWRNG called? through SMC? | 17:57 |
jonwil | Is Linux the only OS that can run on the TI OMAP? | 17:58 |
jonwil | I doubt it | 17:58 |
jonwil | also, having kernel source and having full docs are 2 different things | 17:58 |
jonwil | especially if kernel source is written in a way to hide as much info as possible | 17:58 |
jonwil | e.g. not giving names to hardware registers (or giving useless names) | 17:59 |
freemangordon | Pali: hmm, what we need is smc calls for n900 | 18:00 |
freemangordon | only 3 functions are in trm, with their numbers not matching what is in n900 | 18:01 |
Pali | freemangordon: yes hwrng is used by calling smc | 18:01 |
freemangordon | Pali: the only think we can use is 2.6.28 kernel | 18:01 |
freemangordon | *thing | 18:01 |
Pali | freemangordon: hwrng driver is already in lkml | 18:02 |
freemangordon | I know | 18:02 |
Pali | I sent it for upstream | 18:02 |
freemangordon | I know that too. the point is that our only source of info re ROM on n900 is 2.6.28 | 18:02 |
Pali | but I do not understand why there is rng driver for omap2 and omap4 which using normal memory access | 18:03 |
Pali | but nothing for omap3 | 18:03 |
freemangordon | Pali: where? | 18:03 |
freemangordon | upstream> | 18:03 |
Pali | yes in upstream | 18:03 |
freemangordon | s/>/?/ | 18:03 |
infobot | freemangordon meant: upstream? | 18:03 |
Pali | but for omap3 is only via smc... | 18:03 |
freemangordon | Pali: did you try upstream on n900? | 18:04 |
Pali | and aes/sha/md5 driver using also direct mem access, but secure side needs to be configured | 18:04 |
freemangordon | Pali: hmm, wait. I bet there is an smc call to use crypto module | 18:05 |
Pali | for me it looks like that smc code is used when secure side is not configured for direct mem access | 18:05 |
freemangordon | :nod: | 18:05 |
Pali | no crypto module using direct mem access | 18:05 |
freemangordon | that is why my bet | 18:05 |
Pali | no smc calls | 18:05 |
freemangordon | how do you kno0w that? | 18:05 |
Pali | yes I already looked at crypto module source code | 18:05 |
freemangordon | Pali: that's not relevant | 18:06 |
Pali | it is in upstream kernel | 18:06 |
freemangordon | Pali: isn't it the same for HWRNG? ;) | 18:06 |
freemangordon | upstream uses memory, we SMC | 18:06 |
freemangordon | toldya, I bet it is the same for crypto | 18:07 |
Pali | I do not understand why: hwrng for omap2 and omap4 (both different code), crypto modules (sha, aes, md5) using direct mem read/write | 18:07 |
Pali | and our omap3 hwrng calling smc | 18:07 |
freemangordon | what other device but n900 uses HWRNG from the kernel? | 18:07 |
Pali | and there is no omap3 hwrng in upstream | 18:07 |
Pali | do not know | 18:08 |
freemangordon | N9? | 18:08 |
freemangordon | what about it? merlin1991 ^^^ | 18:08 |
Pali | do not know, need to look into harmattan kernel | 18:08 |
freemangordon | Pali: hmm, I guess we can find the info from harm kernel | 18:09 |
freemangordon | I'll join #harmattan and ask | 18:09 |
Pali | ok | 18:09 |
Pali | if that channel is not dead | 18:09 |
freemangordon | feel free to join :) | 18:09 |
freemangordon | it is not dead | 18:10 |
Pali | freemangordon: look: http://www.droid-developers.org/wiki/Trust_Zone | 18:11 |
freemangordon | Pali: I see nothing of interes here, do I miss something? | 18:16 |
freemangordon | Pali: where is HWRNG driver in KP? | 18:21 |
Pali | in drivers/char/hwrng | 18:22 |
*** joshgillies has quit IRC | 18:25 | |
*** jonwil_ has joined #maemo-ssu | 18:34 | |
*** jonwil has quit IRC | 18:35 | |
*** jonwil_ is now known as jonwil | 18:35 | |
*** jonwil has quit IRC | 18:40 | |
*** Martix has quit IRC | 18:50 | |
*** LauRoman has joined #maemo-ssu | 18:50 | |
FatPhil | jonwil: was yesterday's dose of cosmetics acceptible? | 18:55 |
*** NIN101 has joined #maemo-ssu | 19:05 | |
*** trx has joined #maemo-ssu | 19:10 | |
*** trx has quit IRC | 19:11 | |
*** trx has joined #maemo-ssu | 19:12 | |
*** trx has joined #maemo-ssu | 19:12 | |
FatPhil | fairly subtle passive-aggressive mail composed, ready for sending | 19:35 |
FatPhil | "given that you've shat on our heads, any advice on how we should comb our hair?" | 19:36 |
freemangordon | :D | 19:36 |
freemangordon | Pali: NOLO calls PPA with some other functions, like 37, 29, 21, 20, 19 15 | 19:44 |
freemangordon | ome of them have traces like "Initializing secure storage" (20) "Initializing RPC" (21) | 19:47 |
merlin1991 | FatPhil: did that go to some mailing list, ... can I read it? | 19:47 |
Pali | secure storage?? | 19:48 |
freemangordon | NFC what that is | 19:48 |
Pali | DRM? | 19:48 |
merlin1991 | it is a 2337 haxx0r proof memory location ofc ;) | 19:48 |
Pali | but maemo has no DRM support... | 19:48 |
freemangordon | 18 is "Importing PAPUB keys" | 19:49 |
Pali | PAPUB key is some part of x-loader signing chain maybe... | 19:49 |
Pali | or maybe R&D certificate | 19:49 |
freemangordon | could be | 19:49 |
freemangordon | the point is that we don;t know all PPA functions. and there is a hope there is some function in PPA, which allows us to enable crypto | 19:50 |
FatPhil | merlin1991: *that* didn't go to any mailing list - I was just playing things for laughs ;-) | 19:51 |
Pali | freemangordon: in notebook hdd I have that xloader with crypto enabled functions | 19:52 |
Pali | xloader is for hw rev 2101 only | 19:52 |
Pali | and nolo secondary too | 19:52 |
Pali | can you do some diff against pr1.3 xloader & nolo? | 19:53 |
freemangordon | Pali: there is xloader (some 15k binary) for every rev | 19:54 |
freemangordon | unpack fiasko | 19:54 |
freemangordon | you'll see | 19:54 |
Pali | but not xloader with crypto functions enabled | 19:54 |
freemangordon | I don;t have it | 19:54 |
Pali | that xloader was part of that email sent to maemo ml | 19:54 |
freemangordon | link? | 19:54 |
Pali | uploaded to megaupload | 19:55 |
Pali | so link is dead now | 19:55 |
freemangordon | jacekowski? | 19:55 |
Pali | but I should have copy in notebook hdd | 19:55 |
merlin1991 | Pali: do you have some backup of the notebook hdd somewhere? | 19:55 |
Pali | http://maemo.org/community/maemo-developers/n900_aes_and_sha1-md5_hw_acceleration_drivers/ | 19:55 |
merlin1991 | I'm starting to feel like a lot of important (for us) stuff resides there | 19:56 |
Pali | merlin1991: hdd is near me :-) but I do not have any reader of notebook hdd | 19:56 |
merlin1991 | shouldn't it be regular sata? (if it is not that old) | 19:56 |
Pali | I do not have any machine here... | 19:56 |
Pali | I will never give hdd with notebook to some warranty service provider | 19:57 |
Pali | I do not trust any company... | 19:57 |
* merlin1991 feels like sending his old notebook to Pali, only problem I don't think it is sata yet :D | 19:57 | |
Pali | freemangordon: look at link ^^^ | 19:58 |
freemangordon | "Invalid or Deleted File." :( | 19:58 |
Pali | I got email from HP that I will receive email about case :D | 19:58 |
freemangordon | Pali: don;t tell me you sent your laptop without have your HDD backed up | 19:59 |
Pali | freemangordon: read ^^ I sent it without HDD | 20:00 |
*** dos1 has joined #maemo-ssu | 20:00 | |
freemangordon | Pali: oh, ok :) | 20:09 |
FatPhil | freemangordon: merlin1991: Pali: http://lkml.org/lkml/2013/10/11/533 | 20:15 |
FatPhil | ^^^ tsc2005 mail | 20:16 |
*** sixwheeledbeast has quit IRC | 20:28 | |
freemangordon | good | 20:29 |
*** BCMM has joined #maemo-ssu | 21:01 | |
*** luf has quit IRC | 21:02 | |
*** BCMM has quit IRC | 21:11 | |
*** Woody14619 has joined #maemo-ssu | 21:11 | |
*** Woody14619 has joined #maemo-ssu | 21:11 | |
*** BCMM has joined #maemo-ssu | 21:13 | |
*** BCMM has quit IRC | 22:22 | |
*** BCMM has joined #maemo-ssu | 22:24 | |
*** Vlad_on_the_road has joined #maemo-ssu | 22:29 | |
*** _rd has joined #maemo-ssu | 23:07 | |
*** dos1 has quit IRC | 23:15 | |
*** dafox has joined #maemo-ssu | 23:15 | |
*** dos11 has joined #maemo-ssu | 23:15 | |
*** DrCode has quit IRC | 23:24 | |
*** DrCode has joined #maemo-ssu | 23:28 | |
*** _rd has quit IRC | 23:35 | |
*** Martix has joined #maemo-ssu | 23:41 | |
*** Martix has quit IRC | 23:44 | |
*** Martix has joined #maemo-ssu | 23:45 | |
*** Martix has quit IRC | 23:50 | |
*** Martix has joined #maemo-ssu | 23:51 | |
*** lizardo has quit IRC | 23:57 |
Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!