<RP>
jonmason: I was worried about losing that runner fix so I sent it out
<RP>
rburton: I also sent out those sorting fixes. I suspect there is a bigger underlying issue though
zyga-mbp has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
davidinux has quit [Ping timeout: 252 seconds]
<dagmcr>
Hi guys, yocto-3.1.11 was supposed to be released yesterday but it didn’t. However I can see in dunfell branch 3.1.11 was already bumped earlier this month. Have the plans changed? Or testing went wrong? Thanks!
davidinux has joined #yocto
amitk has joined #yocto
goliath has joined #yocto
<RP>
dagmcr: it is actually in QA atm and due to be released
<dagmcr>
RP: Great! Thanks for the info
<abelloni>
kanavin: it is really weird the test passed
<jonmason>
I'm debating sending patches with a 0/X with a link to the gitlab CI output. Is that worthwhile?
<RP>
jonmason: it always helps to know the testing something has had
<jonmason>
that boost issue on qemuarmv5 is nice to know about :/
<RP>
jonmason: I think mozjs has an issue on v5 too
<jonmason>
double :/
<jonmason>
probably best to open bugs for these, and I'll see about fixing them in the next couple of months
<jonmason>
unless there is more urgency
<RP>
jonmason: no real urgency. I've filed two bugs
<RP>
jonmason: I also filed the poky-tiny arm one
<RP>
again just so we have something to track
<RP>
abelloni: ^^^
Vonter has quit [Ping timeout: 265 seconds]
<jonmason>
yes, I saw the poky tiny email this morning (which is why you don't read work email in bed on a weekend)
<RP>
jonmason: sorry!
* RP
really shouldn't be working
<jonmason>
No, don't get me wrong. I was going to work on it today. I'm just saying that I should be reading work email :)
<jonmason>
I'm a clean plate kind of person, and wanted to get this stuff off my plate
<RP>
jonmason: I just worry at things slipping through cracks. I have echos of the os._exit discussion for qemurunner before :/
<jonmason>
Also, I promised someone I was going to do a blog post for arm on setting up gitlab ci. So, I need to do a rough draft of that as well
<RP>
jonmason: I know what you mean about resolving things, I wanted the pseudo issue I was chasing yesterday sorted out
<RP>
my fixes keep not working :/
<jonmason>
RP: no worries. it seems like I'm chasing bugs shaked out by gitlab ci for weeks now
<jonmason>
oh, and I made the fun mistake of accidentally hosing my sstate hdisk mbr and had to rebuild that from scratch this week
<RP>
jonmason: how do you think I feel with the autobuilder!
<RP>
jonmason: wait until people tell you it would be so much nicer if... :)
<RP>
at least sstate is disposable I guess
* RP
needs to clear his sstate disk
<jonmason>
RP: watching that video from LPC has me wanting to add stuff. Maybe with enough poking rburton will do it ;-)
<jonmason>
sstate is disposable, and I was able to steal from the YP sstate for a little speed-up, but it's become obvious how much I depend on it being there for fast builds
<RP>
jonmason: I invalidate mine so often it isn't funny :(
<jonmason>
my nightlys take about an hour to build lots of machines and run testimage on the qemus. These are taking over 12 without an sstate being there
<RP>
It is just the nature of the stuff I work on
<jonmason>
yes, but you have a massive backend to help you. I have a single desktop PC
<jonmason>
and I need to work on getting more of my local systems in a build setup
<RP>
jonmason: I just have single local server I do my dev work on
<jonmason>
oh jeez, then you must be waiting forever
<RP>
jonmason: it does have 80something cores
<RP>
I only ever really get to mix my changes in with other changes on the main AB
<jonmason>
I would be fun to automate the RPC system I have to power-up some systems, then find some way to monitor the job, then shut them down
<RP>
jonmason: trivial to write a plugin to bitbake to do something when the build completes
<RP>
in fact a metadata eventhandler on BuildComplete would be enough
<abelloni>
RP: I had a look back at what was done for mozjs
<abelloni>
they used EXTRA_OECONF_append = " --disable-ion "
<abelloni>
which made it work on armv5
<abelloni>
it was 2 years ago so it may have been fixed since then
<RP>
abelloni: handy to know. We should probably have at least some v5 testing to try and stop regressions going unnoticed
<RP>
abelloni: its lovely to see the swatbot queue clear. Not sure our AB-INT bug count will agree but...
<jonmason>
RP: lol, that was the exact reason why I added it to our CI. Too many unnoticed regressions in qemuarmv5. If you add it, please let me know so I can remove it from ours.
<ant__>
jonmason, I don't have any armv5 toy anymore but I have experienced many boot issues on real hw beautifully ignored by qemu :)
<jonmason>
ant__: that is fair, but difficult to test in CI :)
<ant__>
offhand I remember gcc6>7 transition with align issues :)
<jonmason>
armv5 is so old, I'm not sure I could add any systems to the internal LAVA test framework
<jonmason>
edit, ask for armv5 systems to be added...
<ant__>
ping hrw
<ant__>
^
<jonmason>
is the RaspPi version 1 armv5?
<ant__>
no, is a developer who now has my armv4/5 zaurus
<ant__>
he promised he'll play again with it
<ant__>
bbl, must reboot, bye
ant__ has quit [Remote host closed the connection]
<RP>
jonmason: do your tests cover polkit?
<jonmason>
RP: currently the minimal testimage, trying to expand to everything that is quick (and currently works)
<RP>
jonmason: quick cheat would just be to add boost and mozjs to the list of targets I guess
<jonmason>
RP: are you referring to my CI, or are you going to add qemuarmv5 to AB?
<jonmason>
also, is systemtap broken in riscv or did I just find a bug (its missing from COMPATIBLE_HOST)? I thought it was working earlier in the week
<RP>
jonmason: I don't know any more :)
<RP>
jonmason: we just upgraded the kernel so that probably broke it
perdmann has joined #yocto
<jonmason>
RP: no problem, I'll take a look and see if it's trivial to fix. If so, patch will be sent. If not, I'll open a bug. work for you? :)
<RP>
jonmason: riscv? we don't officially support riscv :(
<jonmason>
I already have that line. So, that is how I know it's missing
<RP>
jonmason: right, but the recipe specifically says it doesn't build on riscv
<jonmason>
ah
<RP>
so that isn't going to work
<jonmason>
I'm seriously down the rabbit hole of trying to cover my butt for graphics issues with qemuarm ;-)
<jonmason>
"Just add a test in CI to make sure it doesn't break"
<RP>
every time I look at some of this code I regret it
<jonmason>
its always a race to get something done so that you can work on something else
<jonmason>
no regrets
<RP>
jonmason: rburton would have a fit if he was looking at this bit of oeqa :/
<jonmason>
RP: so, we show it to him on monday then?
<RP>
jonmason: tempting :)
wwilly has joined #yocto
<RP>
bug 14575 and 14574 are probably nice easy ones if someone fancies helping
gioyik has joined #yocto
manuel_ has quit [Remote host closed the connection]
manuel_ has joined #yocto
<tlwoerner>
jonmason: i think rpi v1 is armv6
* tlwoerner
has several armv5 things being worked on
<tlwoerner>
yes, the original rpi had an ARM1176JZF-S, ARM11 is armv6 architecture
angolini has quit [Quit: Connection closed for inactivity]
xmn has quit [Ping timeout: 252 seconds]
lucaceresoli has joined #yocto
vd has quit [Quit: Client closed]
Vonter has joined #yocto
lucaceresoli has quit [Ping timeout: 265 seconds]
<jonmason>
I'll have to dig through my pile of old arm boards and see if I have anything older than that
gioyik has quit [Ping timeout: 276 seconds]
* RP
has an arvm4 zaurus somewhere
pabigot has quit [Remote host closed the connection]
pabigot has joined #yocto
florian has joined #yocto
prabhakarlad has joined #yocto
florian has quit [Ping timeout: 260 seconds]
camus has quit [Quit: camus]
gioyik has joined #yocto
vd has joined #yocto
gioyik has quit [Ping timeout: 276 seconds]
florian has joined #yocto
florian has quit [Ping timeout: 260 seconds]
sakoman has joined #yocto
gioyik has joined #yocto
gioyik has quit [Ping timeout: 276 seconds]
gioyik has joined #yocto
gioyik has quit [Ping timeout: 276 seconds]
<abelloni>
jonmason: I can probably get some recent armv5 board shipped to you if you plan to add that in a farm
<abelloni>
sam9x60 was released end of 2019
gioyik has joined #yocto
<jonmason>
abelloni: thanks for the offer. I don't think work would allow outside boards to the LAVA farm. If I can get something like LAVA setup here at home, then I might take you up on it.
<Jawsy>
In a yocto build I have pulled in the meta-virtualization layer to get docker running. But with the change in the kernel config, pthread_setschedparam() has started to fail inside a thread. Any ideas? or am I in the wrong place? The meta-virtualization layer has enabled a bunch of cgroup related configs.
locutusofborg has quit [Ping timeout: 265 seconds]