23:13 <calebccff> bamse: as of 5.17 (maybe some RCs?) I've been hitting a new modem crash, on postmarketOS we have a service which runs some QMI commands to initialise the SIM card, it runs at boot at polls for the modem to appear. It seems like as of recently having it send QMI commands as soon as the modem becomes available causes it to crash

23:14 <bamse> calebccff: i.e. it works if you wait a little bit?

23:14 <calebccff> yeah added a 2 second sleep before sending commands and that fixes it

23:15 <bamse> good

23:15 <bamse> i did see a couple of modem crashes when testing some things yesterday, but it wasn't reliably reproducible

23:16 <calebccff> it's not 100% reliable for me, but happens more often than not, maybe 3/4 boots would fail when the service starts

23:48 <bamse> calebccff: and v5.16 worked fine?

23:49 <calebccff> yeah, no issues there

23:49 <bamse> hmm, nothing obvious in the git log of qrtr, rpmsg or remoteproc...

23:50 <calebccff> hmm, thanks for taking a look. I'll have a go at bisecting

23:50 <bamse> that would be much appreciated

23:50 <bamse> although it really stinks when it's not 100% reproducibl

23:54 <calebccff> yeahh definitely... Hoping to get some board automation hooked up in the future so bisecting and regression testing can get a bit easier

23:55 <bamse> git bisect run with proper board automation is soo nice

23:57 <calebccff> aha I can dream :P

23:58 <calebccff> On the plus side.. OnePlus did make it oh so easy to hook up their devices