Monday, 2023-08-14

*** sakoman <sakoman!> has quit IRC (Quit: Leaving.)00:48
*** belgianguy <belgianguy!> has joined #yocto01:15
*** Ablu <Ablu!~Ablu@user/Ablu> has quit IRC (Ping timeout: 248 seconds)01:17
*** Ablu <Ablu!~Ablu@user/Ablu> has joined #yocto01:18
*** belgianguy <belgianguy!> has quit IRC (Quit: Leaving)01:22
*** brrm <brrm!> has quit IRC (Ping timeout: 246 seconds)01:25
*** brrm <brrm!> has joined #yocto01:26
*** brrm <brrm!> has quit IRC (Excess Flood)01:28
*** brrm <brrm!> has joined #yocto01:28
*** brrm <brrm!> has quit IRC (Ping timeout: 248 seconds)02:09
*** brrm <brrm!> has joined #yocto02:14
*** brrm <brrm!> has quit IRC (Excess Flood)02:15
*** kpo <kpo!> has quit IRC (Ping timeout: 246 seconds)02:17
*** brrm <brrm!> has joined #yocto02:21
*** camus <camus!~Instantbi@> has joined #yocto02:23
*** brrrm <brrrm!> has joined #yocto02:26
*** brrm <brrm!> has quit IRC (Ping timeout: 246 seconds)02:26
*** sakoman <sakoman!> has joined #yocto02:34
*** dash_hope <dash_hope!~dash_hope@> has quit IRC (Ping timeout: 246 seconds)02:41
*** pbsds <pbsds!~pbsds@> has quit IRC (Quit: The Lounge -
*** pbsds <pbsds!~pbsds@> has joined #yocto02:45
*** Circuitsoft <Circuitsoft!> has quit IRC (Quit: Connection closed for inactivity)02:51
*** brrrm <brrrm!> has quit IRC (Ping timeout: 248 seconds)02:52
*** brrm <brrm!> has joined #yocto03:27
*** brrm <brrm!> has quit IRC (Ping timeout: 260 seconds)03:56
*** brrm <brrm!> has joined #yocto03:57
*** brrm <brrm!> has quit IRC (Ping timeout: 246 seconds)04:17
*** brrm <brrm!> has joined #yocto04:18
*** Wouter0100670440 <Wouter0100670440!> has quit IRC (Quit: The Lounge -
*** Wouter0100670440 <Wouter0100670440!> has joined #yocto04:27
*** brrm <brrm!> has quit IRC (Ping timeout: 248 seconds)04:29
khemold_boy:  yocto uses recipe specific sysroots which means it only stages whats needed for a package to build so first thing to check is if you have the needed packages in DEPENDS04:31
khemwhich provide these .pc files04:31
*** brrm <brrm!> has joined #yocto04:33
old_boykhem: thanks. let me check.04:42
*** brrm <brrm!> has quit IRC (Ping timeout: 246 seconds)04:47
*** brrm <brrm!> has joined #yocto04:50
*** brrm <brrm!> has quit IRC (Excess Flood)04:52
*** brrm <brrm!> has joined #yocto04:52
*** brrm <brrm!> has quit IRC (Ping timeout: 240 seconds)04:56
*** brrm <brrm!> has joined #yocto04:57
*** brrm <brrm!> has quit IRC (Excess Flood)05:00
*** amitk_ <amitk_!~amit@> has quit IRC (Ping timeout: 245 seconds)05:09
*** sakoman <sakoman!> has quit IRC (Quit: Leaving.)05:11
*** brrm <brrm!> has joined #yocto05:13
*** goliath <goliath!~goliath@user/goliath> has joined #yocto05:15
*** brrm <brrm!> has quit IRC (Excess Flood)05:15
*** brrm <brrm!> has joined #yocto05:15
*** brrm <brrm!> has quit IRC (Excess Flood)05:18
*** brrm <brrm!> has joined #yocto05:21
*** brrm <brrm!> has quit IRC (Ping timeout: 256 seconds)05:25
*** brrm <brrm!> has joined #yocto05:27
*** brrm <brrm!> has quit IRC (Excess Flood)05:28
*** brrm <brrm!> has joined #yocto05:29
*** brrm <brrm!> has quit IRC (Excess Flood)05:31
*** brrm <brrm!> has joined #yocto05:31
*** brrm <brrm!> has quit IRC (Excess Flood)05:35
*** brrm <brrm!> has joined #yocto05:39
*** brrm <brrm!> has quit IRC (Ping timeout: 246 seconds)05:43
*** brrm <brrm!> has joined #yocto05:45
*** Chaser <Chaser!~Chaser@user/chaser> has joined #yocto06:03
*** brrm <brrm!> has quit IRC (Ping timeout: 245 seconds)06:19
*** mvlad <mvlad!~mvlad@2a02:2f08:470d:cf00:7656:3cff:fe3f:7ce9> has joined #yocto06:19
*** florian <florian!> has joined #yocto06:23
*** brrm <brrm!> has joined #yocto06:43
*** brrm <brrm!> has quit IRC (Excess Flood)06:46
LetoThe2ndyo dudX06:53
*** brrm <brrm!> has joined #yocto06:55
*** brrm <brrm!> has quit IRC (Excess Flood)06:57
*** brrm <brrm!> has joined #yocto06:57
*** zpfvo <zpfvo!> has joined #yocto06:57
*** Guest55 <Guest55!~Guest55@> has joined #yocto06:59
*** Guest55 <Guest55!~Guest55@> has left #yocto07:01
*** frieder <frieder!> has joined #yocto07:08
*** brrm <brrm!> has quit IRC (Ping timeout: 252 seconds)07:17
*** brrm <brrm!> has joined #yocto07:18
*** Chaser <Chaser!~Chaser@user/chaser> has quit IRC (Quit: Chaser)07:19
*** brrm <brrm!> has quit IRC (Ping timeout: 248 seconds)07:25
*** brrm <brrm!> has joined #yocto07:39
*** pabigot <pabigot!> has quit IRC (Ping timeout: 245 seconds)07:41
dvergatalRP: can we have a meeting in here at the earliest on wednesday?07:57
dvergatalRP: because today is a day before national day in poland and everything is working like it is...07:59
*** arisut <arisut!~none@gentoo/developer/alicef> has quit IRC (Quit: install gentoo)08:07
*** arisut <arisut!~none@gentoo/developer/alicef> has joined #yocto08:09
*** arisut <arisut!~none@gentoo/developer/alicef> has quit IRC (Remote host closed the connection)08:11
*** arisut <arisut!~none@gentoo/developer/alicef> has joined #yocto08:12
*** arisut <arisut!~none@gentoo/developer/alicef> has quit IRC (Remote host closed the connection)08:12
*** Chaser <Chaser!~Chaser@user/chaser> has joined #yocto08:13
RPdvergatal: I'm around most days here. Not sure I can add much though, I don't understand that failure either08:14
*** arisut <arisut!~none@gentoo/developer/alicef> has joined #yocto08:15
*** florian <florian!> has quit IRC (Ping timeout: 250 seconds)08:17
*** brrm <brrm!> has quit IRC (Ping timeout: 244 seconds)08:19
*** brrm <brrm!> has joined #yocto08:19
*** brrm <brrm!> has quit IRC (Ping timeout: 246 seconds)08:24
*** brrm <brrm!> has joined #yocto08:35
LetoThe2ndrburton: is falling over for me: Unable to get checksum for linux-yocto SRC_URI entry 0001-Revert-arm64-defc08:36
LetoThe2ndonfig-Enable-Tegra-MGBE-driver.patch: file could not be found08:36
*** brrm <brrm!> has quit IRC (Ping timeout: 244 seconds)08:39
dvergatalRP: ok so I will talk to you on wendnesday08:40
*** brrm <brrm!> has joined #yocto08:41
*** TroubledGuest <TroubledGuest!~TroubledG@> has joined #yocto08:41
*** florian <florian!> has joined #yocto08:45
*** brrm <brrm!> has quit IRC (Ping timeout: 246 seconds)08:45
KanjiMonsterLetoThe2nd: that confuses me a bit, as only adds an append for 6.1, so it shouldn't try to add it for 5.1508:46
KanjiMonsterah moved it to 5.1508:48
*** Chaser_ <Chaser_!~Chaser@user/chaser> has joined #yocto08:50
*** zpfvo <zpfvo!> has quit IRC (Ping timeout: 245 seconds)08:52
KanjiMonsterLetoThe2nd: the issue is likely that ARMFILESPATHS := "${THISDIR}/files:", but the patch(es) are in files/aarch64, that's why it can't find them ( @ jonmason )08:52
*** Chaser <Chaser!~Chaser@user/chaser> has quit IRC (Ping timeout: 256 seconds)08:53
dvergatalRP: I will summarize it briefly for you, my suspicion is that if for some reason one poky build uses different revision without these particular patches for ACLs it may change that sstate, which will remove milliseconds from the source code archived in the sstate tar archive, than if another revision will run again with these patches it will not update the sstate...09:04
*** zpfvo <zpfvo!> has joined #yocto09:06
*** Wouter0100670440 <Wouter0100670440!> has quit IRC (Quit: The Lounge -
*** Wouter0100670440 <Wouter0100670440!> has joined #yocto09:12
*** brrm <brrm!> has joined #yocto09:13
RPdvergatal: sstate filenames have hashes in their filenames which are a sum of all inputs. The input would include the way they're created. There should therefore be one set with the ACL patches and one without. Whilst it is possible something went wrong, we're going to need more evidence than just a suspicion :/09:18
RPdvergatal: It is interesting they're all -src packages. It seems more likely some command in the src copying part of the code has some issue with timestamp tuncation, maybe host specific09:20
rburtonLetoThe2nd: I’m on holiday, tell jonmason. Hopefully our CI exploded over the weekend too.09:26
LetoThe2ndrburton: k, sorry for the noise. :-(09:27
rburtonNo problem, my fault for opening irc. I don’t get notifications on09:28
rburtonMy phone09:28
LetoThe2nd:-) go away! you are not here!09:28
RPrburton: alexandre said it was all the fault of your patches fwiw ;-)09:31
*** brrm <brrm!> has quit IRC (Ping timeout: 250 seconds)09:31
*** brrm <brrm!> has joined #yocto09:32
*** brrm <brrm!> has quit IRC (Excess Flood)09:35
LetoThe2ndRP: a neighbour here has a vw touran, in the "cross" edition, which means there is a "cross touran" print on the side. somebody peeled off the "c", and I just can't unsee it on any "cross" thing now.09:37
JaMazeddii: I see bunch of failures with meta-virtualization and usrmerge, are you already cooking something? I see there is some BIN_PREFIX being used, but not sure what were the use-cases for it as it's set only in,        install -m 755 "${WORKDIR}/k8s-init" "${D}${BIN_PREFIX}/bin" doesn't work well as it doesn't respect ${base_bindir} being /usr/bin with usrmerge09:40
*** brrm <brrm!> has joined #yocto09:41
JaMazeddii: with the same in nerdctl and cloud-init, lxcfs09:42
RPLetoThe2nd: heh, fun :)09:44
*** brrm <brrm!> has quit IRC (Excess Flood)09:45
*** brrm <brrm!> has joined #yocto09:45
*** florian_kc <florian_kc!> has joined #yocto09:46
*** brrrm <brrrm!> has joined #yocto09:50
*** brrm <brrm!> has quit IRC (Read error: Connection reset by peer)09:50
*** brrrm <brrrm!> has quit IRC (Ping timeout: 248 seconds)09:55
*** starblue <starblue!> has quit IRC (Ping timeout: 250 seconds)10:00
*** brrm <brrm!> has joined #yocto10:01
kanavinRP: I can look into stuff this week10:02
*** starblue <starblue!> has joined #yocto10:02
mcfriskparselogs really needs a machine and image specific ways to change the list of accepted error messages. I proposed one solution but that was rejected. the test itself is useful but only after machine/image specific tuning has been done initially.10:15
*** kpo <kpo!> has joined #yocto10:17
*** brrm <brrm!> has quit IRC (Ping timeout: 256 seconds)10:19
mcfriskI have parselogs passing on custom distro and machine on qemu, but need to also support 5 different boards/SoCs from same machine and image, and would thus need to check each of the various failures. managing the exception list without forking poky is really a must.10:19
*** old_boy <old_boy!~old_boy@> has quit IRC (Quit: Client closed)10:19
*** dodoradio <dodoradio!~dodoradio@> has joined #yocto10:21
RPkanavin: My big worry is the 6.4 kernel hangs10:21
*** brrm <brrm!> has joined #yocto10:21
RPkanavin: sadly they're rare. Damaging for CI when they do occur :(10:21
kanavinRP: there needs to be instrumentation for that, unfortunately I'm not a specialist in that area :-/10:24
mcfriskaffecting only Intel x86_64 on qemu?10:26
RPmcfrisk: yes10:27
RPkanavin: what kind of instrumentation are you thinking of?10:28
* mcfrisk is trying to reproduce systemd 254 build failure seen by ci/automation on our side...10:28
kanavinRP: something gdb-like that would probe into a hanging virtual machine to see where it's spending time10:29
mcfriskkanavin: for post boot in runtime tests there are the debug commands which even use qemu debug port to query data10:30
mcfriskif hang happens before serial prompt, then those don't run10:31
RPkanavin: we did have code which was meant to try and help do this kind of thing but it hasn't really worked out so far10:31
RPI'd warn that for a bug like this, the kernel is likely pretty dead so even the output from a connected debugger might not be helpful10:31
mcfriskat least test framework is detecting the hangs and exiting (hopefully killing all processes too)10:31
RP is the more revealing - effectively stack traces from the hung cpus10:33
*** kpo <kpo!> has quit IRC (Ping timeout: 246 seconds)10:35
*** frieder <frieder!> has quit IRC (Read error: Connection reset by peer)10:35
*** brrm <brrm!> has quit IRC (Ping timeout: 246 seconds)10:37
*** frieder <frieder!> has joined #yocto10:43
RPcan't see anything obvious in 6.4.1010:45
*** brrm <brrm!> has joined #yocto10:45
mcfriskmaybe update qemu from 8.0.3 to 8.0.4 ? checking git log..10:49
*** brrm <brrm!> has quit IRC (Ping timeout: 245 seconds)10:50
*** ptsneves <ptsneves!> has joined #yocto10:52
RPmcfrisk: upgrade applies easily enough so probably worth doing...10:54
mcfriskpossibly related ?10:54
*** brrm <brrm!> has joined #yocto10:58
RPmcfrisk: that was the issue we ran into a while ago. It looks similar to this one but the fix does seem to be present in 6.410:59
*** brrm <brrm!> has quit IRC (Excess Flood)11:01
mcfriskRP: ok, qemu git log v8.0.3..v8.0.4 has some possibly relevant commits and fixes for x86_6411:02
JaMazeddii: lxcfs is already fixed in master-next, will send fixes for other 311:06
*** brrm <brrm!~brrm@2a02:8071:b780:ece0::1> has joined #yocto11:09
*** TroubledGuest <TroubledGuest!~TroubledG@> has quit IRC (Quit: Client closed)11:10
*** camus <camus!~Instantbi@> has quit IRC (Ping timeout: 256 seconds)11:11
*** camus <camus!~Instantbi@> has joined #yocto11:12
JaMazeddii: fixes sent, 3rd one (cloud-init) won't be needed if you fix your cloud-init upgrade commit in master-next11:22
*** prabhakarlad <prabhakarlad!> has quit IRC (Quit: Client closed)11:35
*** dodoradio <dodoradio!~dodoradio@> has quit IRC (Quit: dodoradio)11:42
mcfriskRP: would be nice to get into master, I see the same failure with systemd 254.11:47
*** xmn <xmn!~xmn@2600:4040:9390:8c00:6878:433f:857e:5cc8> has joined #yocto11:50
RPmcfrisk: it was ready in the tested queue, merged11:51
*** kpo <kpo!> has joined #yocto11:53
mcfriskRP: thanks!11:54
*** davidinux <davidinux!~davidinux@> has joined #yocto11:59
*** pabigot <pabigot!> has joined #yocto12:11
*** rty <rty!> has joined #yocto12:12
rtyI have multiple packages (say, foo1 and foo2) installing the same files in other package's (say, bar) sysroot (since they are its dependency). how do I specify, that when I build bar, I want to take the files from foo1?12:14
neverpanicI don't think that's possible, unless you remove one of them from the dependency tree, or you change them to no longer conflict (and then maybe use update-alternatives to select one)12:20
mcfriskrty: not easy at all if you always want to build both recipes providing the same files in different binary packages. If you want to decide between two recipes and only build one of them, then it is easier. it is easier to rename the conflicting files to be unique and let things like systemd provide compatibility symlinks.12:20
*** vladest1 <vladest1!~Thunderbi@2a02:1210:760b:9500:b043:a357:8502:83c> has joined #yocto12:28
*** vladest <vladest!> has quit IRC (Read error: Connection reset by peer)12:28
*** vladest1 is now known as vladest12:28
zeddiiJaMa: thanks for the fixes. I'll stack them onto my queued changes, I hopefully will fix one remaining issue shortly and get everything pushed.12:48
*** rty <rty!> has quit IRC (Ping timeout: 246 seconds)13:02
*** rty <rty!> has joined #yocto13:04
JaMazeddii: thanks13:04
rtyit turns out I don't have to solve this, just port against foo2 (there is no hard reason why foo1 is currently required). thanks for replies!13:05
*** olani <olani!> has quit IRC (Remote host closed the connection)13:23
jonmasonLetoThe2nd: why would anyone not want to use the latest kernel?  ;-)13:27
jonmasonproblem reproduced, and should be fixed shortly13:27
*** olani- <olani-!> has quit IRC (Ping timeout: 248 seconds)13:27
*** olani <olani!~olani@> has joined #yocto13:28
jonmasonzeddii: you've been made aware that mickledore is breaking when trying to use xen?  lopper dependency on python3-dtc13:29
zeddiijonmason: ross mentioned it, but it is something you have in your layers when we exchanged messages last. It works fine on my mickledore layers.13:30
jonmasonzeddii: lol, fair enough.  seems like mickledore is going to be a PITA today ;-)13:31
jonmasonis it too late to go back to bed?13:32
zeddiiI can double check, but he said it was a parse time issue ? I was able to build and run mickledore meta-virt xen and lopper. I can fix it easily, but I can't see the breakage on my config.13:32
RPjonmason: was this some kind of setup by ross? :)13:32
zeddiithe move to python3-dtc -> core shouldn't have hit mickledore, and that's the only way that RDEPENDS would have changed (that I can think of), did someone pull in that oe-core commit and match it with your mickledore layes ?13:33
jonmasonRP: Probably13:34
jonmasonzeddii: 100% right, not using the right branch :(13:36
LetoThe2ndjonmason: thankee13:41
LetoThe2ndzeddii: remotely related, any particular reason why docker-compose is using the v2 branch upstream, not v3?13:42
LetoThe2ndjonmason: "mickledore is going to be a PITA today" was my last 2 weeks, basically.13:43
jonmasonLetoThe2nd: just exposing holes in our testing.  It is actually appreciated13:44
LetoThe2ndjonmason: hehe13:45
LetoThe2ndif you're up for a laugh - a BSP i recently encountered had compatibility "gatesgarth honister hardknott langdale" set, and i was like "WUT?"13:46
LetoThe2ndzeddii: and on top of that: 2023-08-14 13:44:46 - ERROR    - ERROR: docker-compose-v2.17.2-r0 do_fetch: Fetcher failure: Unable to find revision 02ad467f89ebc343aa03ce89d18875ea4d604ea3 in branch v2 even from upstream13:47
LetoThe2nd2023-08-14 13:44:46 - ERROR    - ERROR: docker-compose-v2.17.2-r0 do_fetch: Bitbake Fetcher Error: FetchError('Unable to fetch URL from any source.', 'git://;name=compose;branch=v2;protocol=https')13:47
zeddiiLetoThe2nd: wonderful. they deleted branches in the upstream repo.13:52
zeddiieasy enough fix, I'll tweak my master-next commit. but seriously, people need to get their heads on straight. why would someone do that ?13:53
zeddiilooks like they went master->main and nuked the "v2" branch at the same time.13:53
LetoThe2ndzeddii: NFC. and not the RF kind.13:54
qschulzkhem: JaMa: neverpanic: I want to compile qtbase from qt6 with Rockchip libmali.13:55
LetoThe2ndqschulz: i want a sandwich13:55
qschulzlibmali blob is implementing old version of GBM/EGL/GLES/whatever13:55
qschulzto remedy this, I assume, Rockchip created what they call wrappers13:56
qschulzwhich implement the missing functions on top of the old version. They hardcode a bunch of stuff for things to compile and "just work" (maybe?)13:56
qschulzthese wrappers are named after the names of libs you would commonly find in the wild, the proper implementation13:57
*** sakoman <sakoman!> has joined #yocto13:57
qschulze.g.,, etc.13:57
*** olani- <olani-!~olani@> has joined #yocto13:57
qschulzso all programs testing EGL support link against that lib13:57
qschulznow, since it is a wrapper, it doesn't include the symbols from libmali, because why would it need them since they;re in libmal13:58
qschulzhowever, whatever needs to link against (the wrapper) will for sure need to find the symbols from libmali13:58
qschulzI suppose that's what they wanted to do by adding libmali to the NEEDED field of wrappers13:58
qschulzhowever, they aren't automatically pulled in13:58
qschulzwhen you only link against libEGL.so13:59
RPzeddii: it is just a name though and you can just change it...13:59
* zeddii chuckles. 13:59
zeddiiit is definitely Monday.13:59
LetoThe2ndzeddii: such monday, much wow.14:01
qschulznow, there's some weird things going on with linker flags14:02
qschulzIf I don't have -Wl,--as-needed, it finds the symbols in libmali but refuses to link against anyway.14:03
qschulzexcept if I give -Wl,--copy-dt-needed-entries flag14:03
qschulzbut this doesn't work if I also pass -Wl,--as-needed flag14:03
*** prabhakarlad <prabhakarlad!> has joined #yocto14:06
*** paulg <paulg!> has joined #yocto14:09
* RP wonders why the test that fails on all the autobuilders works locally14:10
JPEWqschulz: We use `patchelf --set-soname ${D}${libdir}/` at `do_install` to fix it14:20
JPEWThen symlink all the lib*GL* names it provides to libmali.so14:21
JPEWIt's not "correct" but at least it lets things work without complaining14:21
JPEWOr if you have the version that has the little wrapper libraries instead of symlinks, you still do the --set-soname, but install the wrappers with install_lib14:23
JPEWAh, install_lib is a wrapper we wrote :)14:25
*** Wouter0100670440 <Wouter0100670440!> has quit IRC (Quit: The Lounge -
*** Wouter0100670440 <Wouter0100670440!> has joined #yocto14:32
qschulzJPEW: what does the patchelf command fix as issue? I don't think we've the same issues here?14:42
qschulzJPEW: actually, I just needed to disable the wrappers and the compile script from their meson script just did that by itself14:46
qschulzso, now the issue is that the used version is so old that kmscube refuses to link against it14:46
qschulzand glmark2 refuses to run (which I assume is related too but I don't know)14:46
LetoThe2ndpossibly stupid question - to get started with swupdate on the rpi4, i should be able to use, right? how do I obtain the initial image with A/B partitioning?14:48
*** geoffhp <geoffhp!~GParker@> has joined #yocto14:50
JaMaqschulz: isn't qtbase using pkg-config to find EGL? if so why doesn't it end with both -lEGL and -lmali as your last example?14:52
JaMaJPEW: except when something has in NEEDED and you get file-rdeps QA issue because all symlinks rprovide is libmali.so14:53
qschulzJaMa: nope, not qt6, or at least, not ONLY14:53
*** rsalveti <rsalveti!> has joined #yocto14:53
qschulzI mean... maybe, I don't know what find_library does under the hood14:54
JaMapkg_check_modules(PKG_EGL QUIET egl)14:54
qschulzbut then it tries to compile a small c file by linking against it14:54
qschulzJaMa: ack, so I guess it could be that egl.pc is incorrectly configured somehow?14:54
qschulzi know nothing from pkgconfig so I guess I need to do some reading :)14:55
JaMayes, I would start by checking that14:55
JaMajust try pkg-config --libs egl14:55
JaMain the WORKDIR/devtool of qtbase14:55
qschulzfrom a devshell I guess14:55
JaMas/devtool/devshell/g .. yes14:55
RPsomehow the autobuilder is installing an arm socat package into an x86_64 rootfs and then complaining it doesn't run14:57
JaMaJPEW: I'm working around the evil symlinks with: FILERPROVIDES:/usr/lib/${PN} += ""14:57
neverpanicYeah, patching egl.pc to correctly include libmali is what I'd recommend; you may also have to touch the CMake thing, though.14:57
qschulzJaMa: pkg-config --libs egl14:58
qschulz-lmali -lEGL -ldrm14:58
neverpanicIIRC find_library() does not ask pkg-config unless you specifically use pkg_check_modules() in CMake14:58
JaMait just screws reusing the built binaries on other libEGL implementations (which don't provide libmali as well) as you'll end with both in NEEDED, but that's just the price of evil BSPs14:58
JaMaqschulz: that looks good, now add debug to cmake and check the log.do_configure14:59
JaMa-DCMAKE_FIND_DEBUG_MODE=ON --debug-find --debug-output --trace --trace-expand --check-system-vars --log-level=TRACE --log-context --warn-uninitialized --warn-unused-vars --check-system-vars14:59
JaMafor good read in cold winters..14:59
neverpanicOrder may matter in that response, though. With --as-needed, the compiler may decide that at the point where it sees -lmali on the command line, libmali isn't needed, so it doesn't link it, then it links libEGL, which is needed, but then it fails because libEGL needs symbols from libmali, which aren't anywhere to be found in the rest of the compiler command line.15:00
neverpanicI'd patch that .pc file to list -lmali after -lEGL.15:00
qschulzwill try neverpanic's suggestion first and then JaMa's :)15:02
JaMaisn't there a wrapper around all of these wrappers to wrap it more nicely? :)15:02
qschulzI've read a few things about the check_cxx_source_compiles thing and it seems very limited, so I'm also half wondering if it isn't messing things up as well15:03
*** vmeson <vmeson!> has quit IRC (Quit: Konversation terminated!)15:04
JaMa# Use pkg-config to get the directories and then use these values15:05
JaMa# in the FIND_PATH() and FIND_LIBRARY() calls15:05
JaMa^ seems very limited as well15:05
qschulzneverpanic: mmm qtbase passes the configure step, let's see if it now passes the compile task too :)15:07
qschulz(I have to check I haven't messed things up in other layers/devtool workspaces though before shouting victory :)15:08
*** vvmeson <vvmeson!> has joined #yocto15:09
*** roussinm <roussinm!> has joined #yocto15:11
roussinmI'm trying to add `mold` as a linker available for a sdk, do I need to create mold, mold-cross and mold-crosssdk recipes?15:13
jonmasonLetoThe2nd: mickledore-next should have the fix, if you need it ASAP.  I'm waiting for it to be green before pushing15:15
LetoThe2ndjonmason: thanks! np, not super urgent.15:15
*** florian_kc <florian_kc!> has quit IRC (Ping timeout: 256 seconds)15:18
qschulzneverpanic: nope, didn't work, I forgot to re-enable the wrapper thingy15:22
*** Perflosopher <Perflosopher!~perflosop@> has quit IRC (Quit: Ping timeout (120 seconds))15:24
*** Perflosopher <Perflosopher!~perflosop@> has joined #yocto15:25
RPkanavin: is up for grabs? I think I've an idea on reproducing it...15:27
RP(arm packages creeping into an x86 image test via package feed symlink)15:27
*** rty <rty!> has quit IRC (Quit: Client closed)15:28
JaMaanother 4TB NVME on its way to me for all that sweet sstate :)15:31
*** roussinm <roussinm!> has quit IRC (Quit: WeeChat 3.8)15:31
*** roussinm <roussinm!> has joined #yocto15:32
RPJaMa: OE keeping the tradition of driving hardware sales :D15:35
JaMaand the builds being responsible for global warming :)15:36
RPJaMa: well, hopefully we're also doing our bit with hashequivalence15:37
JaMaluckily gpu miners took the lead in that and we can blame them instead15:37
JaMaunfortunately our builds still have a lot of reproducibility issues and everything is MACHINE_ARCH, so hashequivalence cannot help much in this case15:38
*** frieder <frieder!> has quit IRC (Ping timeout: 245 seconds)15:38
RPJaMa: a shame but hopefully even the core work helps a bit15:40
qschulzJaMa: now wondering what I should be looking for in this huuuuge log.do_configure :D15:41
JaMaqschulz: lEGL :)15:41
JaMathat should filter just reasonable amount of lines15:41
qschulzJaMa: I always have libmali right after (since applying the suggestion from neverpanic)15:42
dvergatalRP: yeah this is really interesting, we need to do than some more testings, because at the moment I have absolutely no idea what to look for, the more that, as I said, I've launched these tests at my machine with newest mint and they all passed me without any problems15:47
*** florian_kc <florian_kc!> has joined #yocto15:48
*** zpfvo <zpfvo!> has quit IRC (Ping timeout: 250 seconds)15:49
*** zpfvo <zpfvo!> has joined #yocto15:50
RPdvergatal: my worry is it depends on the configuration of the host tar :/15:51
*** frieder <frieder!> has joined #yocto15:51
dvergatalRP: hmmmm15:51
dvergatalRP: but I'm setting always --format=posix15:52
*** frieder <frieder!> has quit IRC (Remote host closed the connection)15:52
dvergatalRP: does Alexander tested it on different machine?15:52
RPdvergatal: the autobuilder has many different distros on its workers and we mix them all together, specifically to catch issues like this15:54
dvergatalI'm sorry: did alexander test it on another machine?15:54
dvergatalRP: mhm15:54
*** zpfvo <zpfvo!> has quit IRC (Ping timeout: 246 seconds)15:54
dvergatalgenerally these milliseconds are causing to mutch troubles it would be the best to just remove them15:55
*** zpfvo <zpfvo!> has joined #yocto15:55
RPdvergatal: we'd still need to do that everywhere consistently too though15:55
RPkhem: toolchain tests down to "only" 3261 failures:
RPkhem: mips looking similar numbers to all the other arches15:56
*** Kubu_work <Kubu_work!> has joined #yocto15:57
dvergatalRP: additionally I wouldn't know than from tests if it is posix or gnu tar... which would lead to problems with acls15:58
*** ptsneves <ptsneves!> has quit IRC (Ping timeout: 240 seconds)16:00
*** zpfvo <zpfvo!> has quit IRC (Quit: Leaving.)16:00
khemRP: cool, I saw your patches over the weekend, they looked sane to me. Patching gcc for mips meh :) but alright16:02
*** florian <florian!> has quit IRC (Quit: Ex-Chat)16:07
*** florian_kc <florian_kc!> has quit IRC (Ping timeout: 246 seconds)16:11
dvergatalRP: from what i noticed the zeros are always in reproducibleA which means the format of the tar archive is gnu and should be posix, and looking at the code there is only one possibility at the moment for ipk to be created in gnu format, namely if acl and/or xattr ficher is not enabled16:13
dvergatalRP: are you able to see all files in that build on that machine?16:13
dvergatalRP: by that build I mean this one
dvergatalRP: because if I could only see sstate archives for one of failed packages I would be able to tell you more16:15
*** prabhakarlad <prabhakarlad!> has quit IRC (Quit: Client closed)16:15
*** Guest8563 <Guest8563!> has quit IRC (Ping timeout: 245 seconds)16:23
*** sotaover3ide <sotaover3ide!> has quit IRC (Ping timeout: 246 seconds)16:24
*** olani <olani!~olani@> has quit IRC (Ping timeout: 246 seconds)16:28
*** olani- <olani-!~olani@> has quit IRC (Ping timeout: 260 seconds)16:29
*** Chaser <Chaser!~Chaser@user/chaser> has joined #yocto16:39
*** Chaser_ <Chaser_!~Chaser@user/chaser> has quit IRC (Ping timeout: 245 seconds)16:40
*** prabhakarlad <prabhakarlad!> has joined #yocto16:44
*** prabhakarlad <prabhakarlad!> has quit IRC (Quit: Client closed)16:55
*** belgianguy <belgianguy!> has joined #yocto17:29
belgianguyHi all, I've been progressing quite well with a new attempt, but now I'm facing an issue where the kernel modules are being built, I see a lot of INSTALL steps succeeding, but then in a DEPMOD step, it just fails with a weird eror message "failed to create symbolic link'/lib/modules/"
belgianguyIt's all in mickledore17:39
*** tepperson <tepperson!~tepperson@> has joined #yocto17:41
teppersoni have a recipe that is built with cmake, how do i get the build to give me debug symbols?17:42
*** dodoradio <dodoradio!~dodoradio@> has joined #yocto17:44
*** florian_kc <florian_kc!> has joined #yocto17:56
mischieftepperson: should exist already in the form of splitdebug packages (-dbg), unless you turned it off18:00
teppersonmischief: thanks18:00
*** Chaser <Chaser!~Chaser@user/chaser> has quit IRC (Quit: Chaser)18:11
*** tepperson <tepperson!~tepperson@> has quit IRC (Quit: Client closed)18:11
*** Chaser <Chaser!~Chaser@user/chaser> has joined #yocto18:13
*** pabigot <pabigot!> has quit IRC (Ping timeout: 250 seconds)18:15
*** pabigot <pabigot!> has joined #yocto18:29
*** dodoradio <dodoradio!~dodoradio@> has quit IRC (Quit: dodoradio)18:36
*** olani <olani!> has joined #yocto18:38
*** dodoradio <dodoradio!~dodoradio@> has joined #yocto18:39
*** silurian_invader <silurian_invader!~cambrian_@> has joined #yocto18:59
*** mvlad <mvlad!~mvlad@2a02:2f08:470d:cf00:7656:3cff:fe3f:7ce9> has quit IRC (Remote host closed the connection)19:10
*** vvn <vvn!~vivien@> has quit IRC (Quit: WeeChat 4.0.2)19:13
*** vvn <vvn!~vivien@> has joined #yocto19:19
*** Chaser <Chaser!~Chaser@user/chaser> has quit IRC (Quit: Chaser)19:32
belgianguyI've found the origin of the blacklist line and I see in the generated output the directory contains two subsequent forward slashes in a path, could that be it?19:32
*** silbe <silbe!~silbe@2a03:4000:20:16f:96de:80ff:fe22:1aaa> has quit IRC (Ping timeout: 240 seconds)19:41
*** dodoradio <dodoradio!~dodoradio@> has quit IRC (Quit: dodoradio)19:44
smoogebelgianguy: going from the standard replies from the main heavy hitters in the channel.. any answers would probably be more EU time than now19:49
*** silurian_invader <silurian_invader!~cambrian_@> has quit IRC (Read error: Connection reset by peer)19:50
smoogebelgianguy: that said.. the 99.98 part of the non-existant directory seems the most suspicous to me.. like something was supposed to have a different variable definition but came up with numbers19:50
smoogeI don't know anything about the code involved at all though.. so I am probably missing something obvious19:51
belgianguysmooge, indeed, that jumped out to me as well, I'll try to grep on that bit, maybe I can trace back where the 999.98 comes from :)19:51
*** silbe <silbe!~silbe@2a03:4000:20:16f:96de:80ff:fe22:1aaa> has joined #yocto19:51
*** dodoradio <dodoradio!~dodoradio@> has joined #yocto19:53
belgianguysmooge, found their origin, but not why those numbers are in there19:56
smoogethat sounds like a bad hack to make sure your modules get picked by some other sort19:57
smoogeor a 'this isnt ready to work and I am not sure I want to get bug reports on it'19:57
*** Xagen <Xagen!> has joined #yocto19:58
*** Xagen <Xagen!> has quit IRC (Client Quit)19:58
*** Xagen <Xagen!> has joined #yocto19:59
*** Xagen <Xagen!> has quit IRC (Client Quit)19:59
smoogedifferent codebase but similar code name20:00
smoogeso maybe some code is expecting something else to have made that symlink but its gone?20:02
belgianguysmooge, good find, indeed looks very similar20:03
belgianguysmooge, the code they describe there seems to remove the code with 99.98 in it, and I also get a permission denied error on /lib/modules (as if writing to my own filesystem)20:06
smoogewell I expect it is trying to write to /lib/modules20:09
belgianguysmooge, I think the env var $INSTALL_MOD_PATH is not set:
belgianguywhich maybe rather then writing to a yocto subdir actually tries to write to my host system's /lib/modules20:12
belgianguy(owned by root, which could explain the Permission Denied)20:12
*** Haxxa <Haxxa!> has quit IRC (Quit: Haxxa flies away.)20:15
*** Haxxa <Haxxa!> has joined #yocto20:16
*** silurian_invader <silurian_invader!~cambrian_@> has joined #yocto20:26
*** prabhakarlad <prabhakarlad!> has joined #yocto20:32
*** prabhakarlad <prabhakarlad!> has quit IRC (Quit: Client closed)20:43
*** dodoradio <dodoradio!~dodoradio@> has quit IRC (Quit: dodoradio)20:52
*** belgianguy <belgianguy!> has quit IRC (Quit: Leaving)21:00
*** schtobia <schtobia!> has quit IRC (Quit: Bye!)21:00
*** schtobia <schtobia!> has joined #yocto21:00
*** belgianguy <belgianguy!> has joined #yocto21:08
*** dodoradio <dodoradio!~dodoradio@> has joined #yocto21:16
*** belgianguy <belgianguy!> has quit IRC (Quit: Leaving)21:32
*** belgianguy <belgianguy!> has joined #yocto21:34
*** nerdboy <nerdboy!~nerdboy@gentoo/developer/nerdboy> has quit IRC (Remote host closed the connection)21:47
*** belgianguy <belgianguy!> has quit IRC (Remote host closed the connection)21:50
*** prabhakarlad <prabhakarlad!> has joined #yocto21:52
*** belgianguy <belgianguy!> has joined #yocto22:07
*** florian_kc <florian_kc!> has quit IRC (Ping timeout: 244 seconds)22:16
*** Kubu_work <Kubu_work!> has quit IRC (Quit: Leaving.)22:22
*** Wouter0100670440 <Wouter0100670440!> has quit IRC (Quit: The Lounge -
*** Wouter0100670440 <Wouter0100670440!> has joined #yocto22:27
*** nerdboy <nerdboy!~nerdboy@> has joined #yocto22:36
*** Xagen <Xagen!> has joined #yocto22:56
*** |Xagen <|Xagen!> has joined #yocto22:56
*** camus1 <camus1!~Instantbi@> has joined #yocto22:59
*** camus <camus!~Instantbi@> has quit IRC (Ping timeout: 248 seconds)22:59
*** camus1 is now known as camus22:59
*** Xagen <Xagen!> has quit IRC (Ping timeout: 250 seconds)23:00
*** dodoradio <dodoradio!~dodoradio@> has quit IRC (Quit: dodoradio)23:24
*** goliath <goliath!~goliath@user/goliath> has quit IRC (Quit: SIGSEGV)23:30

Generated by 2.17.2 by Marius Gedminas - find it at!