2019-06-11 15:58:28

by liwei (GF)

Subject: [PATCH] clocksource/arm_arch_timer: mark arch_timer_read_counter() as notrace to avoid deadloop

According to arch_counter_register(), mark arch_counter_get_*() what
arch_timer_read_counter() can be as notrace to avoid deadloop when using
function_graph tracer.

0xffff80028af23250 0xffff000010195e00 sched_clock+64
0xffff80028af23290 0xffff0000101e83ec trace_clock_local+12
0xffff80028af232a0 0xffff00001020e52c function_graph_enter+116
0xffff80028af23300 0xffff00001009af9c prepare_ftrace_return+44
0xffff80028af23320 0xffff00001009b0a8 ftrace_graph_caller+28
0xffff80028af23330 0xffff000010b01918 arch_counter_get_cntvct+16
0xffff80028af23340 0xffff000010195e00 sched_clock+64
0xffff80028af23380 0xffff0000101e83ec trace_clock_local+12
0xffff80028af23390 0xffff00001020e52c function_graph_enter+116
0xffff80028af233f0 0xffff00001009af9c prepare_ftrace_return+44
0xffff80028af23410 0xffff00001009b0a8 ftrace_graph_caller+28
0xffff80028af23420 0xffff000010b01918 arch_counter_get_cntvct+16
...

Fixes: 0ea415390cd3 ("clocksource/arm_arch_timer: Use arch_timer_read_counter to access stable counters")
Signed-off-by: Wei Li <[email protected]>
---
drivers/clocksource/arm_arch_timer.c | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/drivers/clocksource/arm_arch_timer.c b/drivers/clocksource/arm_arch_timer.c
index b2a951a798e2..f4d5bd8fe906 100644
--- a/drivers/clocksource/arm_arch_timer.c
+++ b/drivers/clocksource/arm_arch_timer.c
@@ -149,22 +149,22 @@ u32 arch_timer_reg_read(int access, enum arch_timer_reg reg,
return val;
}

-static u64 arch_counter_get_cntpct_stable(void)
+static u64 notrace arch_counter_get_cntpct_stable(void)
{
return __arch_counter_get_cntpct_stable();
}

-static u64 arch_counter_get_cntpct(void)
+static u64 notrace arch_counter_get_cntpct(void)
{
return __arch_counter_get_cntpct();
}

-static u64 arch_counter_get_cntvct_stable(void)
+static u64 notrace arch_counter_get_cntvct_stable(void)
{
return __arch_counter_get_cntvct_stable();
}

-static u64 arch_counter_get_cntvct(void)
+static u64 notrace arch_counter_get_cntvct(void)
{
return __arch_counter_get_cntvct();
}
@@ -947,7 +947,7 @@ bool arch_timer_evtstrm_available(void)
return cpumask_test_cpu(raw_smp_processor_id(), &evtstrm_available);
}

-static u64 arch_counter_get_cntvct_mem(void)
+static u64 notrace arch_counter_get_cntvct_mem(void)
{
u32 vct_lo, vct_hi, tmp_hi;

--
2.17.1

2019-06-11 16:38:52

by Marc Zyngier

Subject: Re: [PATCH] clocksource/arm_arch_timer: mark arch_timer_read_counter() as notrace to avoid deadloop

Hi Wei,

On 11/06/2019 16:21, Wei Li wrote:
> According to arch_counter_register(), mark arch_counter_get_*() what
> arch_timer_read_counter() can be as notrace to avoid deadloop when using
> function_graph tracer.
>
> 0xffff80028af23250 0xffff000010195e00 sched_clock+64
> 0xffff80028af23290 0xffff0000101e83ec trace_clock_local+12
> 0xffff80028af232a0 0xffff00001020e52c function_graph_enter+116
> 0xffff80028af23300 0xffff00001009af9c prepare_ftrace_return+44
> 0xffff80028af23320 0xffff00001009b0a8 ftrace_graph_caller+28
> 0xffff80028af23330 0xffff000010b01918 arch_counter_get_cntvct+16
> 0xffff80028af23340 0xffff000010195e00 sched_clock+64
> 0xffff80028af23380 0xffff0000101e83ec trace_clock_local+12
> 0xffff80028af23390 0xffff00001020e52c function_graph_enter+116
> 0xffff80028af233f0 0xffff00001009af9c prepare_ftrace_return+44
> 0xffff80028af23410 0xffff00001009b0a8 ftrace_graph_caller+28
> 0xffff80028af23420 0xffff000010b01918 arch_counter_get_cntvct+16
> ...
>
> Fixes: 0ea415390cd3 ("clocksource/arm_arch_timer: Use arch_timer_read_counter to access stable counters")
> Signed-off-by: Wei Li <[email protected]>

There is already a patch[1] queued for that. Hopefully Daniel
and Thomas will get this into mainline shortly.

Thanks,

M.

[1] https://lore.kernel.org/lkml/[email protected]/T/
--
Jazz is not dead. It just smells funny...