Re: [PATCH] rcu/nocb: Add an option to ON/OFF an offloading from RT context

[Uladzislau Rezki]

> On Wed, 11 May 2022 19:17:42 +0200
> Uladzislau Rezki <urezki@xxxxxxxxx> wrote:
> 
> > OK. It was added on the latest kernel:
> > 
> > root@pc638:/home/urezki# cat /sys/kernel/debug/tracing/trace_pipe
> >   vmalloc_test/0-1296    [062] b....    18.157470: 0xffffffffc044e5dc: -> in the local_bh_disable()
> > 
> > root@pc638:/home/urezki# cat /sys/kernel/debug/tracing/trace
> > # tracer: nop
> > #
> > # entries-in-buffer/entries-written: 0/0   #P:64
> > #
> > #                                _-----=> irqs-off/BH-disabled
> > #                               / _----=> need-resched
> > #                              | / _---=> hardirq/softirq
> > #                              || / _--=> preempt-depth
> > #                              ||| / _-=> migrate-disable
> > #                              |||| /     delay
> > #           TASK-PID     CPU#  |||||  TIMESTAMP  FUNCTION
> > #              | |         |   |||||     |         |
> > root@pc638:/home/urezki# uname -a
> > Linux pc638 5.17.0-rc2-next-20220201 #63 SMP PREEMPT Tue May 10 20:39:08 CEST 2022 x86_64 GNU/Linux
> > root@pc638:/home/urezki#
> > 
> > so it shows *bh* disabled sections.
> 
> Ah yes. That was added with commit 289e7b0f7eb47 ("tracing: Account bottom
> half disabled sections."). Maybe I should mark that as stable?
> 
BTW, a commit message of the "289e7b0f7eb47" change exactly describes
my first thought why it behaves like that. So if it is possible it would
be great to have it in the 5.10 and 5.15 stable kernels.

Thanks!

--
Uladzislau Rezki