00:01 <RP> abelloni: it is a bug in our hash equivalence, I was joking since it was just a bit ironic it was your change which triggered us to see this issue

00:15 <paulg> fun fact - the 1st rtc driver was written in response to Alan Cox getting a two socket Pentium 90 (I think?) and implementing the 1st SMP support.

00:16 <paulg> prior to that we just did iopl() and inb/outb from luserspace directly to the rtc hardware - not really SMP safe...

00:17 <RP> paulg: what is wrong with that? :)

00:17 <RP> paulg: also see "leds are simple, they're on or off, how much more to it can there be?"

00:30 <paulg> what is depressing is that old source shows it was 1996 - so that was over a quarter of a century ago. :-/

00:30 <paulg> https://elixir.bootlin.com/linux/v2.6.16-rc2/source/drivers/char/rtc.c

10:08 <wyre> hi guys, can be git tags used to set the PV and PN? 🤔

10:09 <wyre> I was reading this http://www.embeddedlinux.org.cn/OEManual/recipes_versioning.html but maybe is a little outdated

10:31 <Austriker> Hi, I am having a problem with one of my apps that can't link to libomp.so.5. I have added libgomp to RDEPENDS of the recipe but it doesn't do the job. I am trying with DEPENDS. From what I understood OpenMP runtine is in libgomp. My recipe doesn't compile anything just extract binaries and install them. libgomp is already in the IMAGE_INSTALL but

10:31 <Austriker> I don't see in the device libomp.so

12:48 <abelloni> paulg: what is depressing is that it took 13 years to remove it ;)

13:49 <abelloni> thank you ;)

13:53 <paulg> abelloni, deleting stuff can be more difficult than writing it. All the lurkers come crawling out of the woodwork, saying they still use driver XYZ on their old i386sx with less RAM than an average router etc etc.

13:53 <dvorkindmitry> why may I get this message: preferred version 5.4% of linux-at91 not available (for item kernel-module-ip-tables) ?

13:54 <paulg> https://hardware.slashdot.org/story/15/01/22/1419220/user-plea-means-eisa-support-not-removed-from-linux <--- good example

15:18 <JPEW> RP: is it a similar problem we had with -native?

15:18 <JPEW> I guess, native was because we were running programs and this is because of headers, so not quite the same?

15:20 <Austriker> Hi, I am having trouble understanding how to solve this error: QA Issue: -dev package contains non-symlink .so. I have tried adding the so the FILES_${PN} but I get non -dev package containing so. What is the best way to solve this ?

15:20 <marex> is oe (or with poky distro) gcc-sanitizers supposed to generate libasan.a on aarch64 ?

15:21 <marex> Austriker: look at the install .so files , ls -la , that should provide a hint

15:21 <marex> Austriker: basically fix the Makefile install target of the software package

15:23 <wyre> Do I need that https://www.yoctoproject.org/pipermail/yocto/2016-February/028687.html externalgitsrc class to use git tags for PV?

15:23 <Austriker> marex I did the ls -la in the image/usr/lib/ folder and the symlinks seems ok.

15:24 <marex> Austriker: then bitbake -e recipe and look at FILES:${PN} and FILES:${PN}-dev and see what ends where

15:26 <Austriker> marex the -dev package should contain the .so ? maybe I am missing the basics

15:30 <marex> Austriker: no, I dont think it should

15:30 <marex> it never hurts to have a look at one of the other oe-core recipes and compare what they do , when in doubt

17:18 <RP> JPEW: right, similar but different

17:18 <RP> JPEW: the solution may also be similar...

17:22 <JPEW> Ya. I have a few ideas we can think about

17:23 <JPEW> Might be easier to chat about on the Tuesday call?

17:51 <RP> JPEW: That should be fine. I was wondering about doing what we did for natives but for things in the target sysroot

18:45 <marex> RP: hey uh, who is the go-to gcc person /wrt my libasan.a question ?

18:52 <RP> marex: khem or I :/

18:52 <RP> marex: and I don't know

19:06 <marex> RP: well that's fine, I'll keep digging

19:07 <marex> RP: debian has that libasan.a , so I'll reread its gcc build rules

19:40 <Austriker> Hi, I am trying to add a init.d script. I have inherited update-rc.d and I have also added to FILES_PN ${sysconfdir} but when I go on the device the /etc/init.d/ folder doesn't exist. What I am missing ?

19:52 <marex> Austriker: do_install:append() to install the init script into /etc/init.d/ ?

19:53 <marex> Austriker: https://git.openembedded.org/openembedded-core/tree/meta/recipes-connectivity/openssh/openssh_8.8p1.bb#n46 and https://git.openembedded.org/openembedded-core/tree/meta/recipes-connectivity/openssh/openssh_8.8p1.bb#n113

19:54 <marex> git grep init.d and then git grep init in oe-core is a good starting point

19:56 <ziga_> I am trying to create a "makefile" example which is in some folder which is not part of Yocto. I created a recipe and used (a) inherit externalsrc, (b) EXTERNALSRC = <path to sources> (this path has "main.c" and "makefile" inside), (c) EXTERNALSRC_BUILD = <path to build folder>, (d) SRC_URI += "main.c makefile". Then a recipe has a "do_compile(){make}" task and that is it. When I run the recipe it complains: "do_fetch: Fetcher failure: Unable

19:56 <ziga_> to find file file://makefile" I tried removing SRC_URI -- in this case do_compile() runs and prints error: "make: *** No targets specified and no makefile found. Stop." How do I specify sources???

20:28 <Austriker> marex yes I did that also and ssh server is running on the device but /etc/init.d is not present. Is possible that's not there due to systemd ?

23:47 <halstead> The servers that host https://layers.openembedded.org/ and https://swatbot.yoctoproject.org/ rebooted unexpectedly earlier today and services did not come up automatically. Services are online and available now.

23:47 <halstead> I expect that the reboot was due to an automated migration to new physical hardware but I will need to look into it more before I can say for sure.