*** joone has joined #automotive | 00:24 | |
*** mdurnev has joined #automotive | 01:14 | |
*** jlrmagnus has quit IRC | 01:35 | |
*** mvick has quit IRC | 01:36 | |
*** phpOrDie has quit IRC | 02:42 | |
*** phpOrDie has joined #automotive | 03:10 | |
*** phpOrDie has quit IRC | 03:14 | |
*** malditoDev has joined #automotive | 03:14 | |
*** gunnarx has quit IRC | 03:28 | |
*** Sisco has joined #automotive | 03:42 | |
*** aeiche has quit IRC | 04:28 | |
*** mdurnev has quit IRC | 05:03 | |
*** mdurnev1 has joined #automotive | 05:03 | |
*** mdurnev1 is now known as mdurnev | 05:05 | |
*** TwinTurbro has joined #automotive | 05:14 | |
TwinTurbro | wazzap. | 05:15 |
---|---|---|
*** TwinTurbro has left #automotive | 05:31 | |
*** malditoDev has quit IRC | 05:46 | |
*** kanupatar has joined #automotive | 06:35 | |
kanupatar | In my linux R-CarH2 board, when I give the reboot command, it finally goes into halt and not rebooting..any clues? | 06:35 |
*** egy has joined #automotive | 06:36 | |
Figure | morning | 06:38 |
kanupatar | Figure: morning | 06:46 |
*** gmacario has quit IRC | 06:46 | |
kanupatar | see the log of reboot command..Please help me..http://paste.linux.chat/view/91c96051 | 06:48 |
Figure | uh | 06:55 |
Figure | usually reboot is power cable off and on :D | 06:56 |
kanupatar | Figure: :P | 06:56 |
Figure | few days ago I acatually tried reboot on nvidia TK1-PRO board with the nvidia provided image and it also did not work | 06:57 |
Figure | so I guess it's not that trivial to do reboot | 06:57 |
Figure | unfortunately I didn't have time to debug it so I cant help you with this issue either :/ | 06:57 |
*** malditoDev has joined #automotive | 07:11 | |
kanupatar | Figure: yes, most of the arm systems I have shows the same issue\ | 07:13 |
*** kanupatar has quit IRC | 07:48 | |
*** apinheiro has joined #automotive | 08:04 | |
*** egy has quit IRC | 08:21 | |
*** Sisco has quit IRC | 08:30 | |
*** jobol has joined #automotive | 08:30 | |
*** gmacario has joined #automotive | 08:31 | |
*** Tarnyko has joined #automotive | 08:38 | |
*** ctbruce has joined #automotive | 08:38 | |
Tarnyko | morning | 08:41 |
*** toscalix has joined #automotive | 08:45 | |
*** leon-anavi has joined #automotive | 08:48 | |
*** malditoDev has quit IRC | 08:49 | |
*** mdurnev1 has joined #automotive | 09:04 | |
*** mdurnev has quit IRC | 09:07 | |
*** mdurnev1 is now known as mdurnev | 09:07 | |
*** jiada has joined #automotive | 09:24 | |
*** jiada has quit IRC | 09:25 | |
*** jiada has joined #automotive | 09:26 | |
*** jiada1 has joined #automotive | 09:27 | |
*** jonathanmaw has joined #automotive | 09:52 | |
Figure | Tarnyko: nice job with the patches +2 :) | 09:53 |
*** Tarnyko1 has joined #automotive | 10:09 | |
*** Tarnyko has quit IRC | 10:09 | |
Tarnyko1 | Figure: thanks ! we did it at last | 10:09 |
*** egy has joined #automotive | 11:03 | |
*** mdurnev has quit IRC | 12:14 | |
*** fredcadete has quit IRC | 12:34 | |
*** apinheiro has quit IRC | 12:50 | |
*** Sisco has joined #automotive | 13:51 | |
*** egy has quit IRC | 14:03 | |
*** waltminer has joined #automotive | 14:22 | |
*** Sisco has quit IRC | 14:37 | |
*** apinheiro has joined #automotive | 14:50 | |
*** toscalix has quit IRC | 14:51 | |
*** toscalix has joined #automotive | 14:54 | |
*** apinheiro has left #automotive | 15:33 | |
*** alex_h has joined #automotive | 15:38 | |
*** Aleks1917 is now known as Aleks_ | 15:39 | |
*** Aleks_ is now known as Aleks1917 | 15:40 | |
*** n63917 has joined #automotive | 15:57 | |
*** n63917 has left #automotive | 15:58 | |
*** tom__ has joined #automotive | 16:10 | |
*** CTtpollard has quit IRC | 16:10 | |
*** tom__ is now known as CTtpollard | 16:10 | |
*** mvick has joined #automotive | 16:11 | |
waltminer | Tarynyko1 I added you to a few reviews if could take a look at them please | 16:30 |
Tarnyko1 | waltminer, ok, looking right now | 16:40 |
Tarnyko1 | waltminer, done | 16:43 |
waltminer | thank you | 16:43 |
Tarnyko1 | waltminer, yw, by the way we should have a new one for HVAC plugin soon if everything goes right | 16:44 |
*** jlrmagnus has joined #automotive | 16:56 | |
*** egy has joined #automotive | 16:59 | |
*** jlrmagnus has quit IRC | 17:04 | |
*** CTtpollard has quit IRC | 17:16 | |
*** ctbruce has quit IRC | 17:17 | |
*** Tarnyko1 has quit IRC | 17:29 | |
*** jlrmagnus has joined #automotive | 17:34 | |
*** aeiche has joined #automotive | 17:35 | |
jlrmagnus | Sooo. I have a question about root file systems. | 17:44 |
jlrmagnus | In Software Loading Manager, we will transfer the software update bundles as squashfs. | 17:45 |
jlrmagnus | We can then save a bit of disk space by skipping to unpacking step (thus duplicating the data), and simply mount the received image, | 17:45 |
jlrmagnus | So we will have something like /data/sota/updates/update.squashfs which we can mount. | 17:46 |
jlrmagnus | The interesting bit comes when we have a kernel or rootfs update, and need to reboot the IVI into a specific update mode. | 17:47 |
jlrmagnus | The idea is that we, as a part of the update process, reboot to a pre-installed update partition. | 17:47 |
jlrmagnus | The pre-installed partition has a minimalistic root file system and then mounts /data/sota/updates. | 17:49 |
paulsherwood | rjek: ^^ or any other baserockers? we do this already i think... | 17:49 |
paulsherwood | jlrmagnus: trouble is it's already beer oclock in uk | 17:49 |
jlrmagnus | The squasfs file system image in update.squashfs is then mounted, and executed. | 17:49 |
jlrmagnus | paulsherwood, Good: I need the beer creativity to shoot down this scenario. | 17:50 |
paulsherwood | :) | 17:50 |
jlrmagnus | The end game is that the update-mode will have access to the files in the OTA-transmitted squasfs image and can use those to reflash the root partition. | 17:51 |
* paulsherwood is unclear how many folks use irc at the pub | 17:51 | |
jlrmagnus | Not too many. | 17:51 |
jlrmagnus | I guess. | 17:51 |
*** leon-anavi has quit IRC | 17:52 | |
paulsherwood | i do know some... rjek is one for sure :) | 17:52 |
jlrmagnus | A part of this, I guess, is that I just have to type it out to clarify things for my self. | 17:52 |
paulsherwood | sure... keep going | 17:52 |
jlrmagnus | So the update boot gets a bootparam with the path to a squashfs to mount, loopback, once it has mounted /data/sota/updates. | 17:53 |
jlrmagnus | We also set the bootparam init=/mnt/squasfs/do_update.sh | 17:53 |
jlrmagnus | The initrd rc script will mount all the necessary partitions in order for the init command to resolve and start executing. | 17:54 |
jlrmagnus | Soo. In order. | 17:54 |
jlrmagnus | 1. Reboot with a new partition | 17:54 |
jlrmagnus | 2. initrd mounts /mnt/sota/updates | 17:54 |
jlrmagnus | 3. initrd mounts /mnt/sota/updates/update.squashfs as /mnt/squashfs | 17:54 |
jlrmagnus | 4. Init hands over to the regular boot process. | 17:55 |
jlrmagnus | s/init/kernel/ | 17:55 |
jlrmagnus | 5. Init is launched as /mnt/squashfs/do_update.sh, as specified by bootparams. | 17:55 |
jlrmagnus | 6. do_update.sh dd:s out the new root partition using an image stored under /mnt/squashfs. | 17:56 |
jlrmagnus | 7. Unmount. | 17:56 |
jlrmagnus | 8. sync;sync;sync | 17:56 |
jlrmagnus | 9. reboot | 17:56 |
jlrmagnus | 10. Continue SWM update process (which is covered by the spec) | 17:56 |
jlrmagnus | Yes/No? | 17:57 |
paulsherwood | hmmm... i'm not the expert, but this sounds more messy than it needs to be | 17:57 |
jlrmagnus | Ok. The one requirement I have is that we don't have to unpack the blob handed to us from the SOTA client. | 17:57 |
jlrmagnus | Hence the nested mounts. | 17:57 |
paulsherwood | why do you have that requirement? | 17:58 |
persia | I would probably try to do the dd in the initramfs, without a pivot_root. | 17:58 |
paulsherwood | i assume you have other requirements, also? :) | 17:58 |
jlrmagnus | Because tar xf /mnt/sota/updates/update.tgz will double the disk space usage. | 17:58 |
jlrmagnus | That will be a problem when we transfer images. | 17:59 |
jlrmagnus | s/images/partition images/ | 17:59 |
paulsherwood | jlrmagnus: what about making filesystem-level snapshots work, along the lines of btrfs? | 17:59 |
jlrmagnus | persia: Problem with that is that we cannot customize the install scripts. | 17:59 |
*** toscalix has quit IRC | 17:59 | |
jlrmagnus | persia, Correction: I see the deal. | 17:59 |
jlrmagnus | Agree. | 17:59 |
persia | jlrmagnus: Why not? Presumably you have changed the kernel command line, so you use new initramfs, ... | 17:59 |
jlrmagnus | So we would mount and then look for do_update.sh from within the initrd script. | 18:00 |
jlrmagnus | persia, You are right. | 18:00 |
persia | Right. Saves the two-step boot, which should be both faster and safer. | 18:00 |
jlrmagnus | Agree. | 18:00 |
paulsherwood | also, we've been looking hard at how to do updates without rebooting... is there a *need* to reboot in the usecases you're considering, jlrmagnus ? | 18:00 |
persia | (because the update FS is never mounted RW) | 18:00 |
jlrmagnus | paulsherwood, This is a special case. We can still run partition ops, package installs, and external module reflash without rebooting. | 18:01 |
*** egy has quit IRC | 18:01 | |
jlrmagnus | However, sometimes it may be necessary. | 18:01 |
paulsherwood | jlrmagnus: ack | 18:01 |
jlrmagnus | paulsherwood, Checking btrfs | 18:01 |
jlrmagnus | Hmm. A meeting suddenly sprung on me. Back in 10 min. | 18:02 |
paulsherwood | jlrmagnus: tl; dr - you can use snapshots to represent pre-state and desired post-state, flip between them and back. but btrfs specifically would probably not be trustable for automotive production | 18:03 |
*** jonathanmaw has quit IRC | 18:03 | |
*** gmacario has quit IRC | 18:03 | |
*** jlrmagnus has quit IRC | 18:06 | |
*** jlrmagnus has joined #automotive | 18:22 | |
*** waltminer has quit IRC | 18:44 | |
jlrmagnus | Ok. | 18:46 |
jlrmagnus | I'll start off with what we sort of agreed to here and send out another rev of the spec. | 18:46 |
paulsherwood | ack | 18:46 |
jlrmagnus | Thanks for your input, all./ | 18:47 |
paulsherwood | jlrmagnus: you might be interested in https://github.com/CodethinkLabs/live-atomic-update | 18:47 |
jlrmagnus | Checking | 18:47 |
paulsherwood | it's not production-ready, but there has been interest in this in the systemd community iiuc | 18:48 |
*** jlrmagnus has quit IRC | 18:53 | |
*** jlrmagnus has joined #automotive | 19:06 | |
jlrmagnus | Sigh. Lost connection and some entries. | 19:07 |
jlrmagnus | Ahem. https://youtu.be/96UzSHyp0F8?t=186 | 19:08 |
paulsherwood | heh. let's skip all this linux stuff and do everythign in erlang, then :) | 19:09 |
jlrmagnus | Yes! | 19:09 |
jlrmagnus | Me and a few erlang colleagues have actually discussed (when drunk) an erlux distro that ran the erlang vm as a kernel process. | 19:09 |
paulsherwood | jlrmagnus: if you mean you've lost some of the irc history, https://irclogs.baserock.org/automotive/%23automotive.2015-12-11.log.html | 19:10 |
jlrmagnus | That would lead to all kinds of weird but interesting side effects. | 19:10 |
jlrmagnus | Yes. I checked there to see what happened. | 19:10 |
jlrmagnus | Let me dig into the work you've done with containers and pivot_root. For now I need to keep the review process going. | 19:10 |
paulsherwood | jlrmagnus: erlux.org is available!!!!! | 19:10 |
jlrmagnus | Yes! | 19:10 |
jlrmagnus | The future is here! | 19:10 |
*** gunnarx has joined #automotive | 19:11 | |
*** gunnarx has joined #automotive | 19:11 | |
paulsherwood | go for it! | 19:11 |
* paulsherwood gets back to writing a proposal | 19:11 | |
jlrmagnus | Will do. Let me quit JLR, Genivi, and RVI, to focus on a pie-in-the-sky project with a 0.01% chance of being accepted. | 19:11 |
jlrmagnus | paulsherwood, to wrap things up: I will see how your atomic update work fits into the Software Management scheme of things. | 19:12 |
jlrmagnus | gunnarx, Good evening. | 19:12 |
gunnarx | Good evening. Sleep/wake cycle somewhat screwed up here :) | 19:13 |
jlrmagnus | Sorry about that. | 19:13 |
jlrmagnus | Just wrapping up my reply to David's Software Management review email. | 19:14 |
gunnarx | it happens, been a bit under the weather to say the least so I've stayed in bed most of today also | 19:14 |
jlrmagnus | paulsherwood, persia, and I figured out a way to handle a software update boot mode. | 19:14 |
jlrmagnus | gunnarx, I know how we Swedes operate. You've already checked out for the year, haven't you? | 19:14 |
gunnarx | nice one. and here I am feeling sorry for myself not in the mood for jokes | 19:15 |
jlrmagnus | Sorry. I shall be somber. | 19:16 |
gunnarx | I only wish I could check out :) | 19:16 |
jlrmagnus | I hear you. | 19:16 |
jlrmagnus | Are you going to CES? | 19:17 |
gunnarx | the jury is still out | 19:17 |
jlrmagnus | I am actually not going. We are swarming CES this year, so we decided to keep me out of it. | 19:18 |
gunnarx | I see CTtpollard is offline but still congrats on fixing the weirdness on the minnowboard build. Runs green now like the others. | 19:26 |
jeremiah | What does the last S stand for in CES? | 19:36 |
gunnarx | Show | 19:38 |
gunnarx | I see what you did there... | 19:40 |
jeremiah | o_O | 19:40 |
gunnarx | you're implying real engineers don't do "shows" :) | 19:40 |
jeremiah | Heavens no. | 19:40 |
jeremiah | The tech vaudville show must go on | 19:40 |
jeremiah | CES, FOSDEM, same shit | 19:41 |
gunnarx | There's no I in team and there's no Show in FOSDEM. Famous sayings. | 19:41 |
jeremiah | lol | 19:41 |
jeremiah | But there is a me in team | 19:41 |
gunnarx | jlrmagnus, sorry that I've not been in shape to even look at the SW mgmt stuff this week. | 19:47 |
jeremiah | Wow, that pivot_root stuff is interesting. | 19:47 |
gunnarx | I trust it's all good. :) Or we'll know next week anyway. | 19:47 |
jeremiah | jlrmagnus: There is an Erlang kernel: Ling | 19:50 |
jeremiah | There is also a Haskell kernel and an Ocaml kernel: http://wiki.xenproject.org/wiki/Unikernels | 19:50 |
jlrmagnus | Back | 19:51 |
jeremiah | Sheesh, you missed a lot | 19:51 |
jlrmagnus | gunnarx, No problem. | 19:51 |
jlrmagnus | jeremiah, I sent you an email earlier this week on git branching strategies. | 19:52 |
jlrmagnus | i.e. bump. | 19:52 |
jeremiah | Really? How did I miss that? | 19:52 |
* jeremiah looks in his inbox | 19:52 | |
jlrmagnus | jlrmagnus, looks in his outbox | 19:52 |
jlrmagnus | Shit. I never sent it. Sorry. | 19:53 |
jlrmagnus | jlrmagnus, looks in his drafts | 19:53 |
jeremiah | phew | 19:53 |
jeremiah | Although I did get: Software Management - Update bundles | 19:53 |
jeremiah | But we're now talking git | 19:54 |
jlrmagnus | Yes. | 19:54 |
jlrmagnus | We are talking git, and branching, and releases. | 19:54 |
jeremiah | Sounds great. | 19:54 |
*** malditoDev has joined #automotive | 19:56 | |
jlrmagnus | Sent. Since it is a bit of an internal question on how to organize the RVI work, I didn't include the public list. | 19:57 |
jlrmagnus | However, feel free to copy the list as you reply. | 19:58 |
*** Chris___ has joined #automotive | 20:02 | |
*** waltminer has joined #automotive | 20:06 | |
jeremiah | Nice, thanks. | 20:11 |
jeremiah | jlrmagnus: What does Lilli's hook do? | 20:14 |
jeremiah | https://github.com/lillialexis/common/blob/master/commit-msg | 20:14 |
jeremiah | It looks like it just puts the branch name in the commit message? | 20:14 |
Chris___ | Yury of Microchip today reported that the agl-demo-platform build breaks (due to missing AMB dependencies) when updating meta-agl and meta-agl-demo to HEAD. Anyone aware of this? | 20:15 |
*** waltminer has quit IRC | 20:38 | |
jlrmagnus | jeremiah, It just prefixes all commits with the branch that it is commited to. | 20:59 |
jlrmagnus | Makes tracing a lot easier. | 21:00 |
jlrmagnus | gunnarx, The time has come. | 21:00 |
jlrmagnus | This weekend I shall sell my XC60 and get a Jag. | 21:01 |
*** Sisco has joined #automotive | 21:21 | |
aeiche | jlrmagnus, which one are you getting? | 21:32 |
jlrmagnus | XF. | 21:32 |
jlrmagnus | That's what I'm leaning towards, anyway. I'll putter around in the white Evoque until the order comes through. | 21:32 |
*** jobol has quit IRC | 21:56 | |
*** Sisco has quit IRC | 21:57 | |
*** waltminer has joined #automotive | 22:15 | |
*** malditoDev has quit IRC | 22:17 | |
*** malditoDev has joined #automotive | 22:17 | |
*** jlrmagnus has quit IRC | 22:20 | |
*** malditoDev has quit IRC | 22:23 | |
*** Chris___ has quit IRC | 22:31 | |
*** gunnarx has quit IRC | 22:48 | |
*** jlrmagnus has joined #automotive | 23:02 | |
*** gunnarx has joined #automotive | 23:03 | |
*** gunnarx has joined #automotive | 23:03 | |
*** aeiche has quit IRC | 23:07 | |
*** jlrmagnus has quit IRC | 23:33 | |
*** malditoDev has joined #automotive | 23:34 | |
*** jlrmagnus has joined #automotive | 23:52 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!