Hi Paul, On Sat, Feb 29, 2020 at 8:22 PM Paul E. McKenney <paulmck@xxxxxxxxxx> wrote: > In the above stack trace, the stalled CPU (that is, CPU 0) is in the idle > loop, correct? Then it took an interrupt and was processing softirqs > upon return from that interrupt when it took a scheduling clock interrupt, > which did the self-reported stall warning. > > Is this analysis correct? Yes, I think so. > If so, and assuming that this is reproducible, I suggest building your > kernel with CONFIG_RCU_EQS_DEBUG=y, which enables additional debugging > that might be helpful with this issue. Will ask the reporter to try it. I haven't been able to reproduce it myself. Thanks
