On 2019-02-12 18:14:14 [+0100], Frederic Weisbecker wrote: > __local_bh_disable_ip() is neither for strict internal use nor does it > require the caller to disable hardirqs. Probaby a celebration for ancient
Probaby > behaviour. I think the point was to override the IP for the tracer. So everyone else used local_bh_disable() and was recorded as the caller except for softirq.c internal usage where __do_softirq() did also "local_bh_disable()" but recorded its caller (instead recording __do_softirq()). Sebastian