=== chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk [07:35] Bug #1650671 opened: Content sharing from snap common is broken === __chip__ is now known as Chipaca [07:50] zyga: have you looked at snapd#3499 at all? (is that a "you" pr? or is it one for jamie? or...?) [07:50] PR snapd#3499: Spi patch === chihchun_afk is now known as chihchun [07:54] mvo: +1 to merge https://github.com/snapcore/snapd/pull/3464 [07:54] PR snapd#3464: interfaces: put base policy fragments inside each interface [07:56] zyga: did one of the reviewer check each individual snippet that its the same as in the old big table? [07:56] zyga: other than that, +1 [07:57] mvo: I cannot say but since jdstrand reviewed it I really trust it is [07:57] mvo: if you want just go over each patch (they are structured for thaT) [07:57] mvo: so that you see the same stuff in green and red [07:59] PR snapd#3464 closed: interfaces: put base policy fragments inside each interface [08:04] ppisati, mind taking a look at bug 1701018 ? (your opinion would be helpful there i think) [08:04] Bug #1701018: Splash screen is not enabled in kernel [08:07] ogra_: yep, as far as i know when you enable the spashscreen you have to select a bitmap [08:07] and it defaults to the penguin ... [08:07] ogra_: yep [08:08] the prob is that they currently use our linux-generic-raspi2 snap ... so they will likely have to rebuild [08:08] (which will make their security story a bit awkward) [08:08] can you mention that on the bug ? [08:09] ogra_: yeah. go ahead [08:11] ppisati, i was hoping you can do that (i want a more authoritative person speaking than me :) ) [08:12] ogra_: oh k, no prob [08:12] thanks :) [08:38] PR snapd#3541 opened: snapd: fix for snapctl get panic on null config values [08:38] ogra_: replied, but actually his request make sense - he wants to setup the splash screen from uboot and then make the kernel pick it up without touching, i wonder if someone already did it [08:39] ogra_: but they probably did per-build, i doubt there's something generic [08:41] yeah [08:42] ppisati, what i tried here was to set "stdout=" in u-boot (i.e. drop serial and vga) but that just ends up with the kernel blanking the console (and not showing any further output) instead of leaving it untouched [08:42] so i guess you would have to actually turn off the console completely in the kernel and have it not take over hdmi at all [08:43] ... pprob is ... they do digital signage ... so they will need the display later from userspace [08:44] so they want the splashscreen till userspace picks up, then they'll mess with fb from there [08:44] uhm [08:45] well, either that or have the kernel show the same splash when it inits the screen [08:45] they simply want a splash during the whole boot til userspace ... i guess they dont care how they get there [08:45] k [08:45] but no, no dynamic splash support in the kernel [08:46] yeah ... and plymouth is really not an option for the generic initrds [08:46] they might be able to buuld their own ... and end up with a 20MB initrd or so [08:46] i wonder if there's a way to embed a bitmap in DT so both uboot and kernel can be teached to use whatever is there [08:46] *build [08:46] that would generic enough [08:47] oh [08:47] now thats an idea [08:47] but then, we need first to find if they want binary data in DT [08:47] and that will probably blow the space reserved for DT at boot [08:48] in rpi2 at least [08:48] yeah [08:48] but thinking about it ... android also manages to have a bootloader splash that stays [08:48] until surfaceflinger starts in userspace [08:49] though they usually dont have any framebuffer console at all [08:49] but they build the kernel fr that device, so they can hardcode the bitmap there [08:50] ogra_: have you tried asking this question to the graphics people? perhaps they know a way to skip fb initialization if it's already setup [08:51] i'm not sure they build the kernel ... afaik they use our snap and want to get the security benefits from it [08:52] ogra_: they can use our src tree, and just apply a single patch (config + bitmap on top of it), and every time we cut a new release, they rebase and rebuild [08:56] * ogra_ finds psplash ... [08:58] i wonder if we could ship that in our initrd ... looks a lot easier than plymouth [08:58] http://git.yoctoproject.org/cgit/cgit.cgi/psplash/tree/ [09:00] PSplash is a userspace graphical boot splash screen for mainly [09:00] embedded Linux devices supporting a 16bpp or 32bpp framebuffer. It has [09:00] few dependencies (just libc), supports basic images and text and handles [09:00] rotation. [09:01] ogra_: but don't they have to add their logo to initrd though? [09:06] yeah [09:09] PR snapd#3542 opened: cmd,client,daemon: expose "force devmode" in sysinfo === Sir_Gallantmon is now known as Son_Goku [10:39] jdstrand, do we have any interface that allows direct access to /dev/fb0 (i suspect i know the answer, but thought i'd ask anyway) ... [10:40] i just makes a psplash snap that runs fine in devmode ... [10:45] Bug #1699504 changed: "network" plug does not allow outbound ping [10:56] kyrofa, http://paste.ubuntu.com/24980409/ [11:03] kyrofa, looks like your devices uses at91-sama5d3_xplained.dtb ... which isnt in our kernel [11:03] kyrofa, https://goembed.in/2016/03/18/build-and-install-linux-for-arm-cortex-a5-based-embedded-board/ looks like a useful starting page [11:05] ogra_, ah, thank you! How did you find the at91 bit? [11:05] google ;) [11:05] seraching for "ATSAMA5D36 dtb" [11:22] ogra_: yes, framebuffer [11:25] mvo (cc zyga): I reviewed every line of https://github.com/snapcore/snapd/pull/3464 and I used the debug functionaility to compare outputs before and after [11:25] PR snapd#3464: interfaces: put base policy fragments inside each interface === chihchun is now known as chihchun_afk [12:02] hey guys [12:07] jdstrand, bah ... the tool i use uses "dup()" ... [12:08] jdstrand, that is why framebuffer doesnt work ... still blocked by seccomp [12:08] do you think that could be allowed ? [12:12] HAH ... [12:13] cheating helps ... if i use docker-support i get access to the dup syscall [12:13] :D [12:14] niemeyer, https://paste.ubuntu.com/24980832/ [12:15] cachio: Aw, it removed the disks.. [12:15] yes [12:15] cachio: Can you retry with --debug? [12:18] niemeyer, it the same with -debug [12:19] something is hepening before the machine is ready [12:19] after the step Creating configuration on linode:ubuntu-core-16-64 [12:19] cachio: Ah, of course, sorry :( [12:19] cachio: Hmmm [12:19] it waits until boot [12:19] and then it is removing disks [12:19] it is like the image could not be booted [12:19] cachio: Can you rebuild spread locally and disable the disk removal logic? [12:19] cachio: We want to keep around a machine that failed with that error, just once, so we can hand it to support [12:20] niemeyer, ok sure [12:28] PR snapd#3536 closed: snap-seccomp: deal with mknod on aarch64 in the seccomp tests [12:29] niemeyer, (Spread-04): 45.33.76.13 [12:32] niemeyer, cannot boot linode:ubuntu-core-16-64 (Spread-04): configID not found [12:34] niemeyer, (Spread-68): 50.116.8.205 [12:34] this failed === chihchun_afk is now known as chihchun [12:51] ogra_: dup, dup2 and dup3 are alreadyin the default seccomp template [12:54] niemeyer, (Spread-68): 50.116.8.205 [12:54] PR snapd#3543 opened: osutil: allow building on ppc64 (not LE) [12:54] this last one has the disk and config === chihchun is now known as chihchun_afk === om26er_ is now known as om26er [13:34] Jun 29 12:03:07 Coco-XPS-13-9360 kernel: [12493.683255] audit: type=1400 audit(1498734187.179:530): apparmor="DENIED" operation="open" profile="snap.test-gst-clutter.test-gst-clutter" name="/sys/devices/pci0000:00/0000:00:02.0/vendor" pid=30120 comm="test-gst-clutte" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0 [13:37] https://www.irccloud.com/pastebin/5o8o4aC3/gst-error [13:38] tintou, got it, sorry [13:38] Didn't realize I was offline :) === anewman is now known as anewman|away [13:54] PR snapcraft#1384 closed: Support yaml merge tags === anewman|away is now known as anewman [14:18] PR snapcraft#1378 closed: Integration tests for snap command with target arch === anewman is now known as anewman|away === anewman|away is now known as anewman === chiluk_ is now known as chiluk [15:07] * Chipaca hugs jdstrand [15:12] * jdstrand hugs Chipaca back :) === anewman is now known as anewman|away [15:38] PR snapd#3544 opened: tests: make main/help run on opensuse and fedora again [15:38] fginther: Pharaoh_Atem: ^ [15:38] um [15:38] i meant fgimenez, but he's not here [15:48] PR snapd#3532 closed: many: support snap title as localized/title-cased name === anewman|away is now known as anewman [18:25] sergiusens, kyrofa: is https://bugs.launchpad.net/snapcraft/+bug/1590767 somewhere on the roadmap for the next snapcraft? [18:25] Bug #1590767: Support snap installed completion scripts

[18:26] or is there already some way of passing the completer through to snapd which I'm not aware of? [20:23] stgraber: I've updated the bug status for snapcraft. [20:23] sergiusens: great, thanks! [20:24] stgraber: would be nice for the snapd side of the bug to be updated though, your comment gives me confidence it is already available for all snapd installations though [20:27] sergiusens: yeah, I remember mvo saying it would be in 2.26 and the release notes for 2.26 seem to confirm it [20:27] sergiusens: there may issues though since I doubt anyone actually used it given that snapcraft doesn't support it === JanC_ is now known as JanC [23:28] Could somebody please retry https://travis-ci.org/snapcore/snapd/builds/248376205 ? I don't think the timeout is the fault of my PR.