| === jkridner_ is now known as jkridner | ||
| === Ursinha` is now known as Ursinha | ||
| === Ursinha is now known as Guest90314 | ||
| janimo | lilstevie, does rooting always mean 1) downgrading to a firmware that has a vulnerability if the firmware is newer + 2) applying an old root method? | 07:10 |
|---|---|---|
| twb | janimo: on what hardware? | 07:12 |
| janimo | twb, I am interested in tf101g | 07:12 |
| janimo | but for devices with no unlocked bootloader I guess the same generic methdos should apply? | 07:13 |
| twb | g is the "prime" one? | 07:14 |
| infinity | No, prime is tf201. | 07:16 |
| janimo | the asus site docs say theupdater cannot to downgrades so am not sure how the old root tricks fooled it into it | 07:18 |
| janimo | anyway I was a fool to upgrade to ICS before making sure rooting requires a specific version of the firmware present | 07:19 |
| twb | I deliberately bought the one with a known SBK | 07:20 |
| janimo | twb, I bought the only model still on stock, but before making sure it should be rootable :(. I vaguely remembered all transformers are unlocked or rootable now | 07:30 |
| twb | newer ones allegedly are if you ring up asus and ask for the key | 07:34 |
| twb | In return for which they refuse all warranties for that particular S/N | 07:34 |
| janimo | twb, hmm, you mean I could ring up asus and ask for a secure boot key? Or what key do you mean? I would happily do that if realistic | 08:16 |
| janimo | altough I am not sure tf101g qualifies as new enough, it is still last year's model | 08:17 |
| twb | janimo: that was on the primes, I heard | 08:18 |
| twb | Where "ring up" means some wanktacular web form, I assume | 08:18 |
| janimo | yeah, I eard primes are ulockable indeed | 08:18 |
| janimo | btw the xda and various other web resources related to rooting are the most disorganized fountains of knowledge I ave ever seen | 08:19 |
| janimo | it is relatively little possible information to be conveyed spread across numerous threads and lots of participants | 08:20 |
| twb | Well they are web fora | 08:23 |
| twb | i.e. fugly HTML usenet for people who are too stupid to configure a newsreader | 08:24 |
| Daviey | twb: 'fora'? Shame, you are one of them. | 08:56 |
| twb | My high school did not offer latin, sorry | 09:01 |
| twb | https://en.wiktionary.org/wiki/forum#Noun | 09:02 |
| twb | fowler (1e) makes no comment. | 09:03 |
| twb | Nor does Strunk & White (unsurprisingly, it being quite short) | 09:03 |
| twb | Heh, I just noticed the "usage notes" on that URL come from fowler 2e, that travesty | 09:07 |
| ogra_ | rbasak, hmm, where is that binary header you mention in the review ? the merge page doesnt show anything like that | 09:32 |
| ogra_ | oh, i'm blind, ignore me | 09:32 |
| rbasak | ogra_: np. I'm just doing highbank now. Could you hold off on a merge please? | 09:32 |
| rbasak | ogra_: I'm basing my patch on dannf's. | 09:33 |
| ogra_ | sure, i'll fix the header during merge though, no need to wait for the US to get up :) | 09:33 |
| * ogra_ wonders why no image has the new flash-kernel, but they all still build | 09:34 | |
| ogra_ | looks like we have a to loose dependency chain here, they should fail without it | 09:34 |
| janimo | ogra_, is flash-kernel getting unified with the debian one this cycle? | 10:22 |
| ogra_ | janimo, well, as much as we can at least, depends if we get highbank and armadaxp into debian in time :) | 10:23 |
| ogra_ | for all the other arches yes though | 10:23 |
| janimo | sounds good | 10:23 |
| janimo | ogra_, what about the jasper/liveCd stuff, is that migration started? | 10:23 |
| ogra_ | your hack for updating the bootkloader has to go elsewhere too | 10:23 |
| ogra_ | no, but we plan to do this switch too this cycle | 10:24 |
| NCommander | ogra_: is the hardware database for the new f-k uploaded yet? (I'm not sure which package its in) | 10:40 |
| ogra_ | it is in f-k itself | 10:41 |
| ogra_ | infinity wanted to split it out into an arch all package though | 10:41 |
| ogra_ | NCommander, dannf has done an armadaxp patch and rbasak works on highbank | 10:42 |
| ogra_ | so dont worry, its all taken care of | 10:42 |
| NCommander | ogra_: wait, I thought the entire point of this was so f-k could be synced, yet the HW database is in the new package?! | 10:42 |
| ogra_ | first step: sync and make sure all images work again (preferably by not changing flash-kernel but by adapting other packages to match the new world order) | 10:43 |
| ogra_ | next step: make sure our f-k changes land in debian | 10:44 |
| ogra_ | last step: split out the db | 10:44 |
| NCommander | ogra_: then repeat steps one and two. Doesn't it make mor esense to split out the DB first :-) | 10:44 |
| ogra_ | no idea, i think it makes most sense to discuss it with debian and have them do the split ;) | 10:45 |
| rbasak | /usr/sbin/dpkg-reconfigure: flash-kernel is broken or not fully installed | 10:45 |
| rbasak | but apt-get -f install does nothing and there's no other indication of a problem | 10:45 |
| rbasak | Any hints? | 10:45 |
| rbasak | I need to run, back in an hour :-/ | 10:45 |
| ogra_ | try to dpkg -i the deb | 10:45 |
| NCommander | rbasak: is that with the new f-k? I wasn't aware you were taking the new f-k enablement | 10:46 |
| ogra_ | alöso the main/universe issues arent solved yet, did you install the necessary packages ? | 10:46 |
| ogra_ | NCommander, yes, he works on highbank support | 10:46 |
| NCommander | rbasak: I'd much prefer having a boot menu, but I realize this has become a time sink now, so ATM, let's roll with the bootz method you had, then I'll get it fixed for the precise SRU/updated for quantal | 10:46 |
| === Guest90314 is now known as Ursinha | ||
| rbasak | ogra_: I have a working flash-kernel on highbank now. But one error: http://paste.ubuntu.com/1016144/. This appears to be coming from linux-base because it doesn't recognize flash-kernel as a bootloader. Do we know about this already? Presumably it'll affect all the other boards too? | 13:38 |
| ogra_ | well, flash-kernel isnt a bootloader | 13:39 |
| rbasak | That's fine. But linux-base needs to not complain :) | 13:39 |
| ogra_ | well, linux-base also need to move to main, i dont expect the switch to work without hassle (never did) ;) | 13:41 |
| ogra_ | i'll look into it, atm i'm rather trying to find out why the images dont fail if f-k is missing | 13:42 |
| rbasak | I think this should be a Debian issue as well, which is why I'm confused. | 13:46 |
| rbasak | f-k depends on l-b, but l-b will complain unless grub or lilo or a bunch of other bootloaders is installed | 13:47 |
| ogra_ | did you set Bootloader-sets-root in your db entry ? | 13:47 |
| rbasak | Yes | 13:47 |
| ogra_ | to what ? | 13:47 |
| rbasak | to "yes". | 13:47 |
| rbasak | To override the one supplied by U-Boot, which doesn't know | 13:47 |
| rbasak | Previously flash-kernel could write a kernel cmdline into boot.scr, but now it appears that it can't do that any more | 13:47 |
| rbasak | But Bootloader-sets-root: yes will work for now | 13:47 |
| ogra_ | yes means it creates a file in the initrd which needs to contain a UUID | 13:48 |
| ogra_ | sigh, thats horrible perl in the linux-base postinst | 13:48 |
| rbasak | more specifically it overrides the root= kernel parameter | 13:48 |
| ogra_ | well, yes, if the file exists in the initrd it will use the value in there | 13:49 |
| rbasak | Perhaps I could try setting bootargs to omit root= in boot.scr, then I won't need that. | 13:49 |
| * rbasak tries it | 13:50 | |
| ogra_ | hmm, i wonder where "DebianKernel::BootloaderConfig" comes from | 13:50 |
| ogra_ | that seems to be the bit complaining | 13:51 |
| ogra_ | if (!is_fresh_installation() && version_lessthan($ARGV[1], '2.6.32-18')) { | 13:51 |
| ogra_ | DebianKernel::BootloaderConfig::check($deb_arch); | 13:51 |
| ogra_ | } | 13:51 |
| ogra_ | that seems to be the code snippet that causes the complaint | 13:51 |
| ogra_ | and the real issue seems to be that "my @config_files " doesnt have anything for u-boot at all | 13:55 |
| ogra_ | lool, any idea ? ^^^^ | 13:56 |
| ogra_ | rbasak, hmm, looking closer it doews exactly what its supposed to be, what you see there is a note not an error | 14:02 |
| rbasak | ogra_: it stops the install with priority=critical. Not sure why, since the note is priority=high. | 14:03 |
| ogra_ | well, but we should be able to preseed it differently to quieten it | 14:05 |
| rbasak | Shouldn't it be clever enough to detect flash-kernel and shut up? | 14:07 |
| rbasak | The warning is meaningless when flash-kernel is in use | 14:07 |
| ogra_ | flash-kernel has nothing to do with it (as i said, its not a bootloader) | 14:08 |
| ogra_ | the code looks for known bootloader configs and doesnt know about u-boot | 14:09 |
| ogra_ | if i read it right it should be possible to switch that off by preseeding linux-base/disk-id-convert-auto to flase | 14:09 |
| ogra_ | *false | 14:10 |
| ogra_ | rbasak, do you have a virgin env around you could test that in ? | 14:10 |
| rbasak | I can test a preseed option, yes. | 14:11 |
| ogra_ | or just hack the template of the package :) | 14:11 |
| ogra_ | which i think we should do if the package moves to main, having it check the bootloader configs just for reading out the latest kernel version from /boot doesnt really make sense | 14:12 |
| rbasak | This is stupid though. I know that flash-kernel isn't a bootloader. But when we're using flash-kernel, then that depends on linux-base, and we know that no bootloader is required. So linux-base should not complain. It should detect flash-kernel as an exception to its warning, such that if flash-kernel is present it does not need to check for a bootloader and does not need to warn the user. | 14:12 |
| ogra_ | well, i beg to disagree :) | 14:12 |
| rbasak | The message is "The boot loader configuration for this system was not recognized.". If flash-kernel is present, the boot loader configuration can be recognised. It's recognised as do-not-care. | 14:13 |
| rbasak | OK, well I'll file a Debian bug and see what they say about it. | 14:13 |
| ogra_ | linux-base should learn about u-boot ... not about flash-kernel | 14:13 |
| rbasak | flash-kernel is where u-boot knowledge is kept. | 14:13 |
| ogra_ | not always | 14:14 |
| rbasak | Specifically, in that DB now. How else is linux-base to discover the situation? | 14:14 |
| ogra_ | specifically in debian where you are not required to use it at all for arm images | 14:14 |
| rbasak | If linux-base is taking it upon itself to figure out the situation, it should be able to figure out the situation correctly. | 14:14 |
| ogra_ | right | 14:14 |
| ogra_ | but it looks for bootloaders not for flashing tools | 14:14 |
| rbasak | flash-kernel is no longer a flashing tool | 14:15 |
| ogra_ | thats its purpose | 14:15 |
| ogra_ | just because we abuse it differently doesnt change its purpose in debian | 14:15 |
| ogra_ | what i dont get in the beginning is why linux-base checks that stuff at all | 14:16 |
| ogra_ | its only a tool to get you the latest kernel version in /boot | 14:16 |
| rbasak | just reading up a bit: I'd say that linux-base should be figuring out the situation correctly, so it should be looking for both bootloaders _and_ flashing tools | 14:19 |
| ogra_ | rbasak, given we carry ubuntu changes in linux-base anyway 8and will go on doing so as long as debian ships perf in there) we should just change the preseed default | 14:19 |
| rbasak | I'm testing the preseed as we speak | 14:19 |
| ogra_ | i will upload a change if that works and we should be done | 14:19 |
| ogra_ | we really dont want it to touch any bootloader configs in any ubuntu arch anyway | 14:20 |
| lilstevie | janimo: downgrades are typically disallowed meaning new exploits are required | 14:26 |
| rbasak | ogra_: no luck with the preseed. I tried "linux-base/disk-id-convert-auto boolean false" - was that right? | 14:30 |
| rbasak | ogra_: the flash-kernel change is working though, and ready: https://code.launchpad.net/~racb/ubuntu/quantal/flash-kernel/highbank2/+merge/108173 | 14:34 |
| ogra_ | rbasak, awesome, i'll take care for linux-base, dont worry about it | 14:38 |
| rbasak | thanks ogra_! | 14:38 |
| janimo | lilstevie, I was just looking at 'Wolf'd method' (sounds like some math theorem) on xda-forums which according to some allows ICS downgrades | 14:39 |
| ogra_ | rbasak, could be be a bit more verbose in the README entry ? | 14:40 |
| ogra_ | like "Calxeda highbank systems" or something like that ? | 14:40 |
| rbasak | ogra_: I started off doing that, then I realised that the machine type in the db needed to match /proc/cpuinfo, and I made the README match that too. I'm not sure if they're doing that with the other machines or not? | 14:41 |
| ogra_ | no, i dont think armadaxp actually says "Marvell Armada XP Development Board (Ubuntu)" in cpuinfo :) | 14:41 |
| rbasak | :-) | 14:42 |
| rbasak | Go ahead and change if you like | 14:42 |
| ogra_ | will do | 14:42 |
| lilstevie | janimo: pm me please :) | 14:43 |
| ogra_ | rbasak, merged and uploaded | 15:01 |
| rbasak | thanks ogra_! | 15:02 |
| ogra_ | rbasak, linux-base uploaded with the postinst removed completely ... bug 1006717 is the MIR in case you want to follow it | 15:48 |
| ubot2` | Launchpad bug 1006717 in linux-base "[MIR] linux-base" [Undecided,Incomplete] https://launchpad.net/bugs/1006717 | 15:48 |
| rbasak | thanks ogra_! | 15:49 |
| ogra_ | (and i dont think its worth filing a debian bug for linux-base ... it makes totally non-ubuntuish assumptions all over the place :) ) | 15:49 |
| GrueMaster | ogra_: the "Marvell Armada XP Development Board" in flash-kernel (12.04) was copy/paste from cpuinfo. Just FYI. | 17:28 |
| ogra_ | GrueMaster, well, it doesnt have (Ubuntu) appended in cpuinfo i bet | 17:32 |
| ogra_ | i just think the README entry should get you an idea about the HW | 17:32 |
| ogra_ | and "highbank" is somewhat not telling you anything :) | 17:33 |
| GrueMaster | I couldn't tell you what was in cpuinfo at this point. I'm a bit removed from the project. :P | 17:36 |
| GrueMaster | But iirc, I was the one that added the entry to flash-kernel, and it was a straight copy/paste. | 17:36 |
| GrueMaster | Since cpuinfo is largely just raw text from the arm soc code (i.e. not derived from silicon), it could have been (knowing Marvell). | 17:37 |
| ogra_ | well, the db definitely needs the exact match ... the readme not so much | 17:38 |
| GrueMaster | Yes, agreed. | 17:38 |
| GrueMaster | Anyway, thought I'd add my $.02 worth. Back to my real job. | 17:39 |
| ogra_ | heh, have fun | 17:39 |
| dannf | ogra_: oh.. did i add Ubuntu to the wrong field? | 18:14 |
| * dannf made that change post-test, after merging onto the ubuntu tree - might've screwed that up | 18:15 | |
| dannf | oh, nope - that's in the README. w/ that i was following suit w/ the precise's flash kernel which appeared to mark Ubuntu-only additions w/ (Ubuntu) | 18:16 |
| dannf | but that's not in my debian tere | 18:16 |
| XavierMiller | Leave | 18:55 |
| janimo | lilstevie, managed to downgrade using wolf | 22:04 |
| janimo | wolf's method. great.now on to rooting | 22:04 |
| lilstevie | janimo: :) | 23:33 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!