*** dhbiker has quit IRC | 00:46 | |
*** NishanthMenon has quit IRC | 01:03 | |
*** DrCode has quit IRC | 02:05 | |
*** DrCode has joined #maemo-ssu | 02:06 | |
*** LauRoman has joined #maemo-ssu | 03:13 | |
*** LauRoman|Alt has quit IRC | 03:16 | |
*** LauRoman has quit IRC | 03:35 | |
*** xes has quit IRC | 03:52 | |
*** M4rtinK has quit IRC | 03:59 | |
*** futpib has joined #maemo-ssu | 04:04 | |
*** Milhouse has quit IRC | 07:18 | |
*** sparetire_ has quit IRC | 07:26 | |
*** Milhouse has joined #maemo-ssu | 07:33 | |
*** futpib has quit IRC | 08:26 | |
*** LauRoman|Alt has joined #maemo-ssu | 09:26 | |
*** Pali has joined #maemo-ssu | 09:30 | |
Pali | merlin1991: why are calendar-backend, cityinfo, mce, osso-product-info, pulseaudio-policy-enforcement and telepathy-gabble not included in CSSU-T? | 09:43 |
---|---|---|
Pali | telepathy-gabble fix video calls with gmail web plugin, pulseaudio-policy-enforcement just disable useless logs in syslos and calendar-backend fix bug 12682 | 09:45 |
povbot_ | Bug https://bugs.maemo.org/12682 calendar-backend's ICalConverter class does not export proper iCal data | 09:45 |
*** dhbiker has joined #maemo-ssu | 10:21 | |
*** _rd has joined #maemo-ssu | 10:25 | |
*** Pali has quit IRC | 10:39 | |
*** Pali has joined #maemo-ssu | 10:39 | |
*** _rd has quit IRC | 11:04 | |
*** _rd has joined #maemo-ssu | 11:16 | |
freemangordon | Pali: didn't you receive merlin1991's mail where he asked us to check if anything is missing? | 11:35 |
Pali | do not see it :-( | 11:35 |
Pali | subject/message-id? | 11:35 |
freemangordon | just a second | 11:35 |
freemangordon | subject: "CSSU T10 packages" | 11:36 |
freemangordon | Pali: also, what am y supposed to look for in nishanth's log? | 11:37 |
freemangordon | s/y/I | 11:37 |
Pali | freemangordon: do you have account on that TI forum? | 11:38 |
freemangordon | IIRC yes | 11:38 |
freemangordon | forum? | 11:38 |
Sicelo | freemangordon: quick question -- i was on thumb, then updated to the latest CSSU-T .. now there are no thumb packages on my system .. is this expected? | 11:39 |
freemangordon | Sicelo: no | 11:39 |
freemangordon | Pali: e2e.ti.com is not a forum afaik | 11:39 |
Pali | freemangordon: e2e.ti.com or whatever it is | 11:39 |
Pali | can you ask question how to check at runtime if L3 firewall is enabled for omap sha/aes/md5 crypto driver? | 11:40 |
freemangordon | I guess I can :) | 11:41 |
freemangordon | will do, after having a lunch as I am starving | 11:41 |
Sicelo | i guess i must reinstall thumb again .. otherwise i didn't have any bad effects from this 'downgrade' | 11:42 |
Sicelo | or is there another way to check if i'm still on thumb? | 11:42 |
Pali | ok, here is kernel panic log which is generated when L3 firewall is not enabled on HS device and omap-aes.ko is loaded: http://thread.gmane.org/gmane.linux.ports.arm.omap/108397/ | 11:42 |
Pali | Sicelo: check if is deb package has *thumb* in version string | 11:42 |
Pali | thats all... if package has thumb substring in version, then it is thumb compiled, otherwise not | 11:43 |
freemangordon | Pali: BTW can't you ask on #armlinux? | 12:00 |
Pali | we already asked on linux arm and omap ML ... | 12:01 |
freemangordon | Sicelo: do as Pali said, it makes no sense -testing to overwrite -thumb, unless you have very old version | 12:01 |
freemangordon | Pali: ok | 12:01 |
freemangordon | need to figure out how to ask questions on e2e | 12:01 |
freemangordon | hmm, only omap4 and omap5 forums | 12:04 |
freemangordon | :( | 12:04 |
Sicelo | well, it did, curious | 12:06 |
*** arcean has joined #maemo-ssu | 12:12 | |
freemangordon | Pali: http://e2e.ti.com/support/dsp/omap_applications_processors/f/447/t/391467 | 12:12 |
Pali | freemangordon: thanks! | 12:13 |
sixwheeledbeast | Sicelo: you definitely have no thumb packages at all? | 12:23 |
Sicelo | i know .. that's why i was asking if the update should have done this. i *did* have thumb before updating to cssu-t. of course it's too late now, haha .. will just reinstall thumb sometime soon | 12:35 |
bencoh | what do you mean by "updating to cssu-t" ? | 12:37 |
*** futpib has joined #maemo-ssu | 12:38 | |
*** Pali has quit IRC | 12:55 | |
*** Pali has joined #maemo-ssu | 13:04 | |
freemangordon | Pali: for that aes to function, what was needed - new NOLO or new xloader? | 13:25 |
Pali | both | 13:25 |
freemangordon | hmm, doesn't make sense | 13:25 |
Pali | and even new flasher.rover binary :D | 13:25 |
freemangordon | I know those are packaged together, but is everything different? | 13:26 |
Pali | yes, both nolo and xloader are different | 13:26 |
Pali | I can send binaries via emails | 13:26 |
freemangordon | I have them | 13:27 |
freemangordon | RX-51_2009SE_crypto.tar.gz ? | 13:27 |
freemangordon | Pali: did you try to flash only xloader or only nolo? | 13:28 |
Pali | I have n900-crypto.tar.gz, omap-hw.tar.gz, nolo.tar.gz, omap-sha1-md5-standalone.diff, omap-aes-standalone.diff | 13:29 |
Pali | I did not tried to flash that images yet | 13:29 |
freemangordon | Pali: my idea is - find how is that FW enabled and where (nolo or xloader), then find a way to do it from the kernel itself | 13:30 |
Pali | but I will try on testing N900 from Doc in January | 13:30 |
freemangordon | I bet there is some SMC that can be (ab)used | 13:30 |
Pali | tmlind suggested to dump some omap regs on normal xloader and that patched xloader if they are different | 13:30 |
freemangordon | BTW ARM hexrays works on both nolo and xloader | 13:30 |
freemangordon | Pali: yep, makes sense | 13:31 |
Pali | maybe it can be used for that runtime detection | 13:31 |
Pali | I was told that new patched xloader (or NOLO??) has linked some binary blob from TI which doing something | 13:31 |
Pali | needed for omap aes | 13:31 |
freemangordon | hmm | 13:32 |
Pali | I think it will be just example blob which enable L3 firewall | 13:32 |
freemangordon | enable? | 13:32 |
freemangordon | or rather disable | 13:32 |
Pali | enable access | 13:33 |
Pali | in public TRM is written something about configuring L3 firewall for GP devices | 13:33 |
freemangordon | which TRM, 35xx? | 13:34 |
Pali | OMAP34xx_ES3.1.x_PUBLIC_TRM_vZT.pdf | 13:35 |
freemangordon | I have ZM here, shouldnt be that much different | 13:35 |
Pali | 5.2 L3 Interconnect | 13:36 |
freemangordon | yeah, I am reading through it | 13:37 |
freemangordon | not that I understand what is written there :D | 13:38 |
Pali | 5.2.5.4.1 Protection Mechanism (PM) Register Description | 13:39 |
Pali | there is L3_PM_REQ_INFO_PERMISSION_i and L3_PM_READ_PERMISSION_i | 13:39 |
Pali | and L3_PM_ADDR_MATCH_k | 13:40 |
Pali | I think some of those registers we should use | 13:40 |
Pali | but I do not understand that documents too :-( | 13:41 |
freemangordon | hmm, xloader and nolo are SMALLER in RX-51_2009SE_crypto.tar.gz than in PR1.3 fiasko | 13:47 |
*** dhbiker has quit IRC | 13:56 | |
Pali | freemangordon: can you tell me sha1sum of RX-51_2009SE_crypto.tar.gz? | 13:59 |
Pali | I have: cf6eac7aed41b5161d50c0361378f6bd1cc06d28 n900-crypto.tar.gz | 14:00 |
*** dhbiker has joined #maemo-ssu | 14:09 | |
freemangordon | Pali: same here | 14:16 |
*** _rd has quit IRC | 14:31 | |
freemangordon | Pali: hmm, I can;t see anything different in xloader, besides date string and some other trivial stuff | 14:32 |
freemangordon | so it must be in nolo | 14:32 |
Pali | really? xloader is same? | 14:37 |
Pali | freemangordon: then we should be able to enable aes from kernel... | 14:38 |
Pali | I still do not know why was needed new flasher binary for flashing... | 14:39 |
Pali | I think it is doing same calls | 14:39 |
Pali | it just extend Mk II flashing protocol which is used in PC-Suite and Update mode | 14:39 |
Pali | but I did not see anything new in normal flashing | 14:40 |
freemangordon | Pali: I guess dmitry just packaged it all together, for consistency | 14:44 |
freemangordon | it was dmitry, correct? | 14:44 |
Pali | yes | 14:45 |
Pali | "In order to use AES hw it is necessary to enable it during boot from Secure side." | 14:45 |
Pali | "It is done by special code which is compiled as part of bootloader." | 14:45 |
Pali | "You need special boot loader, because it has secure code to enable it.." | 14:46 |
Pali | "You cannot do it from non-secure side." | 14:46 |
Pali | "Nolo has binary blob for that." | 14:46 |
freemangordon | see, it is nolo | 14:46 |
Pali | from Dmitry Kasatkin | 14:46 |
*** _rd has joined #maemo-ssu | 14:47 | |
Pali | I think that with NOLO he means both xloader+secondary images | 14:47 |
Pali | but do not know now... | 14:47 |
freemangordon | maybe there is modified PPA in NOLO which enables it | 14:47 |
* freemangordon checks NOLO | 14:48 | |
Pali | is NOLO running as secure or non secure? | 14:48 |
freemangordon | non-secure AFAIK | 14:48 |
freemangordon | but PPA runs in secure mode | 14:48 |
freemangordon | and afaik PPA is loaded by SMC call | 14:48 |
Pali | then only x-loader can set own PPA call | 14:49 |
freemangordon | hmm, no, nolo can do it as well | 14:49 |
freemangordon | it just needs to issue the correct SMC | 14:49 |
Pali | then you should do it also from kernel | 14:49 |
freemangordon | not sure we can do it, as PPA is signed | 14:50 |
freemangordon | but yes, I guess we can load PPA from kernel | 14:50 |
Pali | http://www.droid-developers.org/wiki/Category:Security | 14:50 |
freemangordon | if we have the blob and the signature | 14:50 |
Pali | http://www.droid-developers.org/wiki/Secure_Services | 14:50 |
Pali | looks like some services are SW | 14:51 |
freemangordon | SW? | 14:52 |
Pali | loaded by SW | 14:52 |
freemangordon | yep, my point exactly | 14:52 |
freemangordon | you can load PA by issuing a SMC call | 14:53 |
Pali | then it looks like NOLO loads its own service for HW rng and allow kernel to use it via SMC | 14:53 |
freemangordon | so, maybe it is possible to strip that blob (and signature) from NOLO and load it from kernel | 14:53 |
freemangordon | :nod: | 14:53 |
Pali | possible | 14:54 |
freemangordon | hmm, NOLO is completely different, from the start | 14:56 |
freemangordon | Pali: do you remember at which address is NOLO loaded? | 14:59 |
Pali | maybe 0x80400000 | 15:01 |
freemangordon | thanks | 15:01 |
Pali | it is what I have in unfinished spl port of uboot | 15:02 |
Pali | ~ping | 15:05 |
infobot | ~pong | 15:05 |
*** M4rtinK has joined #maemo-ssu | 15:06 | |
Pali | yes, in uboot is comment that Nokia X-Loader loads secondary image to address 0x80400000 | 15:07 |
Pali | ~ping | 15:07 |
infobot | ~pong | 15:07 |
*** sixwheeledbeast has left #maemo-ssu | 15:38 | |
*** sixwheeledbeast has joined #maemo-ssu | 15:39 | |
*** sixwheeledbeast has left #maemo-ssu | 15:56 | |
*** sixwheeledbeast has joined #maemo-ssu | 16:13 | |
merlin1991 | Pali: I'll prepare a T10.1 in jan that will include the missing bits | 16:14 |
Pali | ok | 16:15 |
merlin1991 | I'd do it this week, but starting with the 27th I'll be gone untill the 1st, don't want to risk it that I botch something and the repo is foobar for so long | 16:18 |
*** _rd has quit IRC | 16:34 | |
freemangordon | Pali: oh, modified nolo seems to set IBE bit in aux CR :) | 16:45 |
*** sparetire_ has joined #maemo-ssu | 17:08 | |
freemangordon | Pali: dammit, I can't find significant differences in NOLO either, besides this IBE bit in ACR, but it should be unrelated :( | 17:47 |
*** futpib has quit IRC | 18:32 | |
*** futpib has joined #maemo-ssu | 18:38 | |
*** _rd has joined #maemo-ssu | 18:48 | |
*** _rd has quit IRC | 18:50 | |
Pali | funny... | 20:09 |
Pali | need to flash all combinations of nolos and xloaders and see what happens... | 20:10 |
*** dhbiker has quit IRC | 21:52 | |
*** nox- has joined #maemo-ssu | 23:17 | |
*** Pali has quit IRC | 23:34 |
Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!