On Fri, 7 Jul 2023 08:37:29 +0300
Dan Carpenter [off-list ref] wrote:

quoted

quoted

This is a sleeping function.  

Hmm, this is an interesting scenario. If this triggers, it means that the
system is likely locked up by the function graph tracer. The only way to
stop the hang, is via calling ftrace_graph_stop(). But you are correct,
that's calling something that can crash the system as well.

If anything, it should be called after the dump_on_oops output, with a
warning to reboot the machine.

IOW, yes, it's doing something buggy, but pretty much the only other
alternative is to call panic(). Not sure that's better :-/

Perhaps the solution is simply to move it to after the dump, with a warning
saying: "Dazed and confused, and trying to continue, but please reboot the machine!"

??  

I feel like sleeping in atomic bugs used to be more of a big deal back
in the day when systems only had one CPU.  In those days it was way more
common for it to lead to a hang, but these days we quite often
re-schedule the sleeping process on a different CPU and recover.  (I
haven't actually looked at how processes are moved to different CPUs
but this is just my theory of why we see fewer real life hangs from this
bug today).

Sleeping while atomic is still a bug. It's just this particular code
path is where I don't know the best way to solve it. It's a start up
test (only enabled on development machines), and when it gets to where
it calls that function that sleeps in atomic, the system is already
hung. That code path detected that the function graph tracer is in some
kind of dead loop (which it may have caused), and it tries to stop that
dead loop by disabling it. But unfortunately, to disable it, it calls a
sleeping function!

Perhaps we just comment it and say, "Yes this is is buggy, but if we
are here, we already hit a bug". ;-)

-- Steve