09:57 <Jones42> OnkelUlla: a colleague recently pointed me towards https://git.yoctoproject.org/meta-security/tree/classes/dm-verity-img.bbclass . So now I'm a bit confused regarding the differences between that one and your proposed bbclass and why you made the effort of writing your own. Could you elaborate a bit?

12:07 <Tyaku> Hello, I have a question that is not directly linked to yocto, but maybe I have more chances here than #linux:

12:07 <Tyaku> On our board we have configured the PMIC DA9062 (watchdog, thermal, rtc, regulators). I know it is working because the board starts with memorised time and we see probe message during the boot. The issue is when we use "reboot" command. In the function 'do_kernel_restart()', the kernel call the callbacks registered by drivers to process the restart. Like all watchog driver, the da9062 have a function

12:07 <Tyaku> registered to do it ('da9062_wdt_restart'). This function try to transmit a message by I2C to tell the DA9062 to shutdown but during the transmission it receive a NACK.

12:08 <Tyaku> My question is maybe simpe for people who know it: The drivers that are started by the Linux based on the DTS are probed when Linux starts, but are they "uninitialized" when linux stop ? (processing the shutdown is maybe the last action)

12:09 <Tyaku> I suspect that the I2C or some clocks or something is uninitialized at the end of the reboot sequence, when we try to send a message on I2C to tell the PMIC to "shutdown" (to process the reboot)

12:09 <Tyaku> https://community.renesas.com/rz/f/rz-forum/34897/rzg2ul---crash-on-reboot---da9062-0-0058-failed-to-shutdown-err--6

12:33 <mbulut> does anyone know how to enable systemd-networkd-persistent-storage.service?

12:54 <mbulut> ok, looks like i need a newer version of systemd than what kirkstone ships... is it even a good idea to backport systemd or should i expect compatibility issues?

14:22 <moto-timo> Anybody know why attr recipe in oe-core doesn't provide /usr/include/attr/xattr.h like https://packages.debian.org/bookworm/amd64/libattr1-dev/filelist

14:28 <moto-timo> (I specifically need it in attr-native)

14:30 <RP> moto-timo: did you enable extended-attr ?

14:34 <moto-timo> RP: how does one enable extended-attr (and no I did not)

14:34 <moto-timo> quick search in docs and poky/meta didn't find anything (and the attr recipe doesn't have a PACKAGECONFIG)

14:36 <RP> moto-timo: xattr in DISTRO_FEATURES iirc, may be the default now

14:36 <moto-timo> RP: ah, just found that in https://git.yoctoproject.org/poky/commit/?id=92fc3ef9738165300a4daed4c90526dbbad43149

14:37 <RP> moto-timo: I was just giving ideas, it may or may not be related

14:37 <moto-timo> RP: at least a thread I can tug on ;) thank you

19:11 <philipl_> Hi, I'm currently looking into setting up a shared sstate mirror and as part of that also read up on hash equivalence. The documentation on hash equivalence says "The shared Hash Equivalence server needs to be maintained together with the Shared State cache. Otherwise, the server could report Shared State hashes that only exist on specific clients." and I don't fully understand why the two need to be maintained together. Assuming that a client

19:11 <philipl_> contributes a task hash to the hash equivalence server while not contributing the corresponding sstate object: Wouldn't the next builder that produces the same outputs for that task pick up this hash and produce the cache object for that unihash?

21:00 <JPEW> philipl_: Ya, but it's not ideal