Re: [PATCH 01/24] arm/stm32f405: Fix realization of "stm32f2xx-adc" devices

["Philippe Mathieu-Daudé" <philmd@redhat.com>]

On 5/27/20 11:27 AM, Markus Armbruster wrote:
> Alistair Francis <alistair23@gmail.com> writes:
> 
>> On Mon, May 18, 2020 at 10:08 PM Markus Armbruster <armbru@redhat.com> wrote:
>>>
>>> Alistair Francis <alistair23@gmail.com> writes:
>>>
>>>> On Sun, May 17, 2020 at 10:06 PM Markus Armbruster <armbru@redhat.com> wrote:
>>>>>
>>>>> stm32f405_soc_initfn() creates six such devices, but
>>>>> stm32f405_soc_realize() realizes only one.  Affects machine
>>>>> netduinoplus2.
>>>>>
>>>>> I wonder how this ever worked.  If the "device becomes real only on
>>>>> realize" thing actually works, then we've always been missing five of
>>>>> six such devices, yet nobody noticed.
>>>>
>>>> I must have just been testing the first ADC.
>>>>
>>>>>
>>>>> Fix stm32f405_soc_realize() to realize all six.  Visible in "info
>>>>> qtree":
>>>>>
>>>>>      bus: main-system-bus
>>>>>        type System
>>>>>        dev: stm32f405-soc, id ""
>>>>>          cpu-type = "cortex-m4-arm-cpu"
>>>>>        dev: stm32f2xx-adc, id ""
>>>>>          gpio-out "sysbus-irq" 1
>>>>>     -    mmio ffffffffffffffff/00000000000000ff
>>>>>     +    mmio 0000000040012000/00000000000000ff
>>>>>        dev: stm32f2xx-adc, id ""
>>>>>          gpio-out "sysbus-irq" 1
>>>>>     -    mmio ffffffffffffffff/00000000000000ff
>>>>>     +    mmio 0000000040012000/00000000000000ff
>>>>>        dev: stm32f2xx-adc, id ""
>>>>>          gpio-out "sysbus-irq" 1
>>>>>     -    mmio ffffffffffffffff/00000000000000ff
>>>>>     +    mmio 0000000040012000/00000000000000ff
>>>>>        dev: stm32f2xx-adc, id ""
>>>>>          gpio-out "sysbus-irq" 1
>>>>>     -    mmio ffffffffffffffff/00000000000000ff
>>>>>     +    mmio 0000000040012000/00000000000000ff
>>>>>        dev: stm32f2xx-adc, id ""
>>>>>          gpio-out "sysbus-irq" 1
>>>>>          mmio 0000000040012000/00000000000000ff
>>>>>        dev: stm32f2xx-adc, id ""
>>>>>          gpio-out "sysbus-irq" 1
>>>>>     -    mmio ffffffffffffffff/00000000000000ff
>>>>>     +    mmio 0000000040012000/00000000000000ff
>>>>>        dev: armv7m, id ""
>>>>>
>>>>> The mmio addresses look suspicious.
>>>>
>>>> Good catch, thanks :)
>>>
>>> I'd love to squash in corrections, but I don't know the correct
>>> addresses.  Can you help?
>>
>> Yep, thanks for squashing it in.
>>
>> The three addresses are:
>>
>> 0x40012000
>> 0x40012100
>> 0x40012200
>>
>> and they all share interrupt number 18.
> 
> An the other three?  There are six devices in total...

Alistair looked at the stm32f205, which has 3 ADCs (and seems correct).

> 
>> Let me know if you want me to do it.

Alistair, the problem is the stm32f405. I guess the bug come from
copy/pasting then modifying for the shared IRQ.

    /* Timer 2 to 5 */
    for (i = 0; i < STM_NUM_TIMERS; i++) {
       ...
    }
    ...
    /* ADC device, the IRQs are ORed together */
    ...
    dev = DEVICE(&(s->adc[i])); <=== i = STM_NUM_TIMERS = 4
    object_property_set_bool(OBJECT(&s->adc[i]),
                             true, "realized", &err);

Only ADC#3 is realized, then mapped at 0x40012000.

You need to unparent the others anyway, so it is easier to realize
adc[0] and unparent the rest:

    for (i = 1; i < STM_NUM_ADCS; i++) {
       ...
    }