On Tue, Apr 30, 2024 at 12:30 PM Eric Dumazet edumazet@google.com wrote:
On Tue, Apr 30, 2024 at 12:17 PM Davide Caratti dcaratti@redhat.com wrote:
hello,
On Tue, Apr 30, 2024 at 12:17 AM Naresh Kamboju naresh.kamboju@linaro.org wrote:
While running selftests: tc-testing: tdc.sh the following kernel warnings, kernel Bug, kernel oops and kernel panic noticed with Linux next-20240429 tag kernel as per the available data.
This build config is from kselftest merge config[1].
Reported-by: Linux Kernel Functional Testing lkft@linaro.org
selftests: tc-testing: tdc.sh log and crash log
the problem is created by [1]. I think that at least we need to guard against failures to allocate sch->cpu_bstats and sch->cpu_qstats, otherwise the dynamic key is registered but never unregistered (though the key is freed in the error path of of qdisc_alloc() ). But there might be also something else; however, I can reproduce some similar splat, will follow-up on the list.
sorry for the noise,
davide
[1] https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next.git/commit/?...
I just had 5 or 6 syzbot reports about this issue.
I tested the following fix.
diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 31dfd6c7405b01e22fe1b8c80944e2bed7d30ddc..d3f6006b563ccd8827b7af362ce9dceaa78f8841 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -982,6 +982,7 @@ struct Qdisc *qdisc_alloc(struct netdev_queue *dev_queue,
return sch;errout1:
lockdep_unregister_key(&sch->root_lock_key); kfree(sch);errout: return ERR_PTR(err);
Davide, I will let you send a formal fix. Let me know if you want me to release a syzbot report.