2019-06-10 16:49:58

by Dmitry Osipenko

Subject: [PATCH v2 6/6] clocksource/drivers/tegra: Restore base address before cleanup

We're adjusting the timer's base for each per-CPU timer to point to the
actual start of the timer since device-tree defines a compound registers
range that includes all of the timers. In this case the original base
need to be restore before calling iounmap to unmap the proper address.

Signed-off-by: Dmitry Osipenko <[email protected]>
---
drivers/clocksource/timer-tegra.c | 2 ++
1 file changed, 2 insertions(+)

diff --git a/drivers/clocksource/timer-tegra.c b/drivers/clocksource/timer-tegra.c
index 2a428fdf702f..7be91db98bd7 100644
--- a/drivers/clocksource/timer-tegra.c
+++ b/drivers/clocksource/timer-tegra.c
@@ -345,6 +345,8 @@ static int __init tegra_init_timer(struct device_node *np, bool tegra20,
irq_dispose_mapping(cpu_to->clkevt.irq);
}
}
+
+ to->of_base.base = timer_reg_base;
out:
timer_of_cleanup(to);

--
2.21.0

2019-06-14 15:49:33

by Jon Hunter

Subject: Re: [PATCH v2 6/6] clocksource/drivers/tegra: Restore base address before cleanup

On 10/06/2019 17:44, Dmitry Osipenko wrote:
> We're adjusting the timer's base for each per-CPU timer to point to the
> actual start of the timer since device-tree defines a compound registers
> range that includes all of the timers. In this case the original base
> need to be restore before calling iounmap to unmap the proper address.
>
> Signed-off-by: Dmitry Osipenko <[email protected]>
> ---
> drivers/clocksource/timer-tegra.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/drivers/clocksource/timer-tegra.c b/drivers/clocksource/timer-tegra.c
> index 2a428fdf702f..7be91db98bd7 100644
> --- a/drivers/clocksource/timer-tegra.c
> +++ b/drivers/clocksource/timer-tegra.c
> @@ -345,6 +345,8 @@ static int __init tegra_init_timer(struct device_node *np, bool tegra20,
> irq_dispose_mapping(cpu_to->clkevt.irq);
> }
> }
> +
> + to->of_base.base = timer_reg_base;
> out:
> timer_of_cleanup(to);

So what you are saying is that because we don't know which CPU executes
the tegra_init_timer() function, then it is necessary to restore the
base. IOW if it is not CPU0, then the base will be updated and hence,
need to be restored. Correct?

Jon

--
nvpublic

2019-06-14 16:53:34

by Dmitry Osipenko

Subject: Re: [PATCH v2 6/6] clocksource/drivers/tegra: Restore base address before cleanup

14.06.2019 18:48, Jon Hunter пишет:
>
> On 10/06/2019 17:44, Dmitry Osipenko wrote:
>> We're adjusting the timer's base for each per-CPU timer to point to the
>> actual start of the timer since device-tree defines a compound registers
>> range that includes all of the timers. In this case the original base
>> need to be restore before calling iounmap to unmap the proper address.
>>
>> Signed-off-by: Dmitry Osipenko <[email protected]>
>> ---
>> drivers/clocksource/timer-tegra.c | 2 ++
>> 1 file changed, 2 insertions(+)
>>
>> diff --git a/drivers/clocksource/timer-tegra.c b/drivers/clocksource/timer-tegra.c
>> index 2a428fdf702f..7be91db98bd7 100644
>> --- a/drivers/clocksource/timer-tegra.c
>> +++ b/drivers/clocksource/timer-tegra.c
>> @@ -345,6 +345,8 @@ static int __init tegra_init_timer(struct device_node *np, bool tegra20,
>> irq_dispose_mapping(cpu_to->clkevt.irq);
>> }
>> }
>> +
>> + to->of_base.base = timer_reg_base;
>> out:
>> timer_of_cleanup(to);
>
> So what you are saying is that because we don't know which CPU executes
> the tegra_init_timer() function, then it is necessary to restore the
> base. IOW if it is not CPU0, then the base will be updated and hence,
> need to be restored. Correct?

We know what the CPU is, it is always CPU0. What we don't know is what TIMER is
assigned for CPU0. On Tegra210 it is TIMER10, for other Tegra's it is TIMER0.