On Wed, Jul 17, 2019 at 11:22:15PM -0700, Paul E. McKenney wrote:
> On Thu, Jul 18, 2019 at 02:43:58PM +0900, Masami Hiramatsu wrote:
> > Remove rcu_read_lock()/rcu_read_unlock() from debug exception
> > handlers since the software breakpoint can be hit on idle task.
Why precisely do we need to elide these? Are we seeing warnings today?
> The exception entry and exit use irq_enter() and irq_exit(), in this
> case, correct? Otherwise RCU will be ignoring this CPU.
This is missing today, which sounds like the underlying bug.
Thanks,
Mark.
>
> Thanx, Paul
>
> > Actually, we don't need it because those handlers run in exception
> > context where the interrupts are disabled. This means those are never
> > preempted.
> >
> > Reported-by: Naresh Kamboju <naresh.kamb...@linaro.org>
> > Cc: Paul E. McKenney <paul...@linux.ibm.com>
> > Signed-off-by: Masami Hiramatsu <mhira...@kernel.org>
> > ---
> > arch/arm64/kernel/debug-monitors.c | 14 ++++++++------
> > 1 file changed, 8 insertions(+), 6 deletions(-)
> >
> > diff --git a/arch/arm64/kernel/debug-monitors.c
> > b/arch/arm64/kernel/debug-monitors.c
> > index f8719bd30850..48222a4760c2 100644
> > --- a/arch/arm64/kernel/debug-monitors.c
> > +++ b/arch/arm64/kernel/debug-monitors.c
> > @@ -207,16 +207,16 @@ static int call_step_hook(struct pt_regs *regs,
> > unsigned int esr)
> >
> > list = user_mode(regs) ? &user_step_hook : &kernel_step_hook;
> >
> > - rcu_read_lock();
> > -
> > + /*
> > + * Since single-step exception disables interrupt, this function is
> > + * entirely not preemptible, and we can use rcu list safely here.
> > + */
> > list_for_each_entry_rcu(hook, list, node) {
> > retval = hook->fn(regs, esr);
> > if (retval == DBG_HOOK_HANDLED)
> > break;
> > }
> >
> > - rcu_read_unlock();
> > -
> > return retval;
> > }
> > NOKPROBE_SYMBOL(call_step_hook);
> > @@ -305,14 +305,16 @@ static int call_break_hook(struct pt_regs *regs,
> > unsigned int esr)
> >
> > list = user_mode(regs) ? &user_break_hook : &kernel_break_hook;
> >
> > - rcu_read_lock();
> > + /*
> > + * Since brk exception disables interrupt, this function is
> > + * entirely not preemptible, and we can use rcu list safely here.
> > + */
> > list_for_each_entry_rcu(hook, list, node) {
> > unsigned int comment = esr & ESR_ELx_BRK64_ISS_COMMENT_MASK;
> >
> > if ((comment & ~hook->mask) == hook->imm)
> > fn = hook->fn;
> > }
> > - rcu_read_unlock();
> >
> > return fn ? fn(regs, esr) : DBG_HOOK_ERROR;
> > }
> >
