Hi Ulf,
On Wed, Jun 29, 2016 at 7:30 PM, Ulf Hansson <ulf.hansson@xxxxxxxxxx> wrote:
> On 28 June 2016 at 18:26, Geert Uytterhoeven <geert@xxxxxxxxxxxxxx> wrote:
>> This patch (commit eb13a0a1b6d5d5c2 in pm/linux-next) breaks resume on
>
> Thanks for reporting!
>
>> sh73a0/kzm9g and r8a73a4/ape6evm. On these boards, the Ethernet controller is a
>> child of a local bus (bsc), whose clock (zb) is controlled through pm_clk and
>> simple-pm-bus, cfr.
>>
>> arch/arm/boot/dts/r8a73a4-ape6evm.dts
>> arch/arm/boot/dts/r8a73a4.dtsi
>> arch/arm/boot/dts/sh73a0-kzm9g.dts
>> arch/arm/boot/dts/sh73a0.dtsi
>>
>> During resume, the bus clock is not enabled, causing an imprecise abort
>> when accessing the Ethernet controller's registers. With some debug code
>> added:
>
> I was looking into this in some more detail.
>
> As you are stating that Ethernet device is a child device to the local
> bus device, I would have expected that the pm_runtime_get_sync()
> invoked during ->probe() in the Ethernet driver should cause its
> parent device (the bus device) to be "forced" resumed.
>
> That's because the pm_runtime_get_sync() should trigger an increased
> usage count of the parent device. Later, when
> pm_runtime_force_resume() validates the count for the bus device, it
> should be 2 and thus it should lead to that it decides to *not* defer
> the resume of the device. Apparently this isn't happening for some
> reason...
>
> Perhaps there's a pm_suspend_ignore_children() set for the parent?
> Or perhaps the parent/child relationship isn't set up correctly?
No, ignore_children = 0. I've dumped all fields from the bus' prepare
callback I had added for Kevin's test:
PM: Suspending system (mem)
simple-pm-bus fec10000.bus: simple_pm_bus_prepare
can_wakeup = 0
async_suspend = 0
is_prepared = 0
is_suspended = 0
is_noirq_suspended = 0
is_late_suspended = 0
early_init = 1
direct_complete = 0
wakeup_path = 0
syscore = 0
no_pm_callbacks = 0
timer_expires = 0
usage_count = 1
child_count = 1
disable_depth = 0
idle_notification = 0
request_pending = 0
deferred_resume = 0
run_wake = 0
runtime_auto = 1
ignore_children = 0
no_callbacks = 0
irq_safe = 0
use_autosuspend = 0
timer_autosuspends = 0
memalloc_noio = 1
request = 0
runtime_status = 0
runtime_error = 0
autosuspend_delay = 0
last_busy = 4294937777
active_jiffies = 0
suspended_jiffies = 430
accounting_timestamp = 4294937777
PM: suspend of devices complete after 4.394 msecs
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@xxxxxxxxxxxxxx
In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds
