On 10/15/2013 02:39 PM, Grant Likely wrote:

All the users of of_irq_parse_raw pass in a raw interrupt specifier from
the device tree and expect it to be returned (possibly modified) in an
of_phandle_args structure. However, the primary function of
of_irq_parse_raw() is to check for translations due to the presence of
one or more interrupt-map properties. The actual placing of the data
into an of_phandle_args structure is trivial. If it is refactored to
accept an of_phandle_args structure directly, then it becomes possible
to consume of_phandle_args from other sources. This is important for an
upcoming patch that allows a device to be connected to more than one
interrupt parent. It also simplifies the code a bit.

The biggest complication with this patch is that the old version works
on the interrupt specifiers in __be32 form, but the of_phandle_args
structure is intended to carry it in the cpu-native version. A bit of
churn was required to make this work. In the end it results in tighter
code, so the churn is worth it.

This commit appears to cause kernel error spew while performing CPU
hotplug operations on Tegra114. For example, see the log at the end of
this message.

I guess this boot-time message is the clue:

[    0.000000] arch_timer: No interrupt available, giving up

which corresponds to this diff in /proc/interrupts:

- 27:       2929       2216       2027       1922       GIC  27  arch_timer

The relevant DT node is probably:

        timer {
                compatible = "arm,armv7-timer";
                interrupts =
                        <GIC_PPI 13
                                (GIC_CPU_MASK_SIMPLE(4) | IRQ_TYPE_LEVEL_LOW)>,
                        <GIC_PPI 14
                                (GIC_CPU_MASK_SIMPLE(4) | IRQ_TYPE_LEVEL_LOW)>,
                        <GIC_PPI 11
                                (GIC_CPU_MASK_SIMPLE(4) | IRQ_TYPE_LEVEL_LOW)>,
                        <GIC_PPI 10
                                (GIC_CPU_MASK_SIMPLE(4) | IRQ_TYPE_LEVEL_LOW)>;
        };

I guess this is a similar problem to that which Olof reported; there's
no reg property in that node, yet #size-cells=<0> since the timer node
is a direct child of the root node.

Hotplug error spew:

root@localhost:~# ./cpuonline.py 
echo 0 > /sys/devices/system/cpu/cpu2/online
[   46.753347] CPU2: shutdown
echo 0 > /sys/devices/system/cpu/cpu1/online
[   48.848007] CPU1: shutdown
echo 0 > /sys/devices/system/cpu/cpu3/online
[   50.897720] CPU3: shutdown
echo 1 > /sys/devices/system/cpu/cpu1/online
[   52.949980] CPU1: Booted secondary processor
[   52.961188] Clockevents: could not switch to one-shot mode: dummy_timer is not functional.
[   52.969566] Could not switch to high resolution mode on CPU 1
echo 1 > /sys/devices/system/cpu/cpu2/online
[   54.999852] CPU2: Booted secondary processor
[   55.011164] Clockevents: could not switch to one-shot mode: dummy_timer is not functional.
[   55.019543] Could not switch to high resolution mode on CPU 2
echo 0 > /sys/devices/system/cpu/cpu1/online
[   57.048730] CPU1: shutdown
echo 1 > /sys/devices/system/cpu/cpu3/online
[   59.088927] CPU3: Booted secondary processor
[   59.101149] Clockevents: could not switch to one-shot mode: dummy_timer is not functional.
[   59.109517] Could not switch to high resolution mode on CPU 3
echo 1 > /sys/devices/system/cpu/cpu1/online
[   61.129737] CPU1: Booted secondary processor
[   61.141161] Clockevents: could not switch to one-shot mode: dummy_timer is not functional.
[   61.149540] Could not switch to high resolution mode on CPU 1