From: Frank Rowand <[email protected]>
Fix kernel build warning:
drivers/of/unittest-data/tests-interrupts.dtsi:32.26-35.6: Warning (interrupt_map): /testcase-data/interrupts/intmap1: Missing '#address-cells' in interrupt-map provider
A recently implemented dtc compiler warning reported the dts problem.
Signed-off-by: Frank Rowand <[email protected]>
---
drivers/of/unittest-data/tests-interrupts.dtsi | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/of/unittest-data/tests-interrupts.dtsi b/drivers/of/unittest-data/tests-interrupts.dtsi
index 9b60a549f502..8c2b91b998aa 100644
--- a/drivers/of/unittest-data/tests-interrupts.dtsi
+++ b/drivers/of/unittest-data/tests-interrupts.dtsi
@@ -31,6 +31,7 @@ test_intmap0: intmap0 {
test_intmap1: intmap1 {
#interrupt-cells = <2>;
+ #address-cells = <1>;
interrupt-map = <0x5000 1 2 &test_intc0 15>;
};
--
Frank Rowand <[email protected]>
On Thu, Oct 28, 2021 at 7:58 PM <[email protected]> wrote:
>
> From: Frank Rowand <[email protected]>
>
> Fix kernel build warning:
> drivers/of/unittest-data/tests-interrupts.dtsi:32.26-35.6: Warning (interrupt_map): /testcase-data/interrupts/intmap1: Missing '#address-cells' in interrupt-map provider
>
> A recently implemented dtc compiler warning reported the dts problem.
>
> Signed-off-by: Frank Rowand <[email protected]>
> ---
> drivers/of/unittest-data/tests-interrupts.dtsi | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/drivers/of/unittest-data/tests-interrupts.dtsi b/drivers/of/unittest-data/tests-interrupts.dtsi
> index 9b60a549f502..8c2b91b998aa 100644
> --- a/drivers/of/unittest-data/tests-interrupts.dtsi
> +++ b/drivers/of/unittest-data/tests-interrupts.dtsi
> @@ -31,6 +31,7 @@ test_intmap0: intmap0 {
>
> test_intmap1: intmap1 {
> #interrupt-cells = <2>;
> + #address-cells = <1>;
Notice that we have 2 nodes with interrupt-map here. One has
'#address-cells' and one doesn't. Why? Because we need to test that
the code can handle both cases.
The dtc warnings are more what should 'new' users do. I don't know
what DTs don't have #address-cells, but my guess is ancient ones.
Rob
On 10/28/21 9:07 PM, Rob Herring wrote:
> On Thu, Oct 28, 2021 at 7:58 PM <[email protected]> wrote:
>>
>> From: Frank Rowand <[email protected]>
>>
>> Fix kernel build warning:
>> drivers/of/unittest-data/tests-interrupts.dtsi:32.26-35.6: Warning (interrupt_map): /testcase-data/interrupts/intmap1: Missing '#address-cells' in interrupt-map provider
>>
>> A recently implemented dtc compiler warning reported the dts problem.
>>
>> Signed-off-by: Frank Rowand <[email protected]>
>> ---
>> drivers/of/unittest-data/tests-interrupts.dtsi | 1 +
>> 1 file changed, 1 insertion(+)
>>
>> diff --git a/drivers/of/unittest-data/tests-interrupts.dtsi b/drivers/of/unittest-data/tests-interrupts.dtsi
>> index 9b60a549f502..8c2b91b998aa 100644
>> --- a/drivers/of/unittest-data/tests-interrupts.dtsi
>> +++ b/drivers/of/unittest-data/tests-interrupts.dtsi
>> @@ -31,6 +31,7 @@ test_intmap0: intmap0 {
>>
>> test_intmap1: intmap1 {
>> #interrupt-cells = <2>;
>> + #address-cells = <1>;
>
> Notice that we have 2 nodes with interrupt-map here. One has
> '#address-cells' and one doesn't. Why? Because we need to test that
> the code can handle both cases.>
> The dtc warnings are more what should 'new' users do. I don't know
> what DTs don't have #address-cells, but my guess is ancient ones.
>
> Rob
>
I had hoped to build all of the .dts files in the Linux tree, with the
new dtc, but did not get to that today. That should flush out any
cases that would result in build fail from the new approach of treating
all warnings as errors. I may get to that tomorrow.
If there any any existing .dts files that will trigger the interrupt
map warning, will we require that they be fixed so that the build will
not fail?
-Frank
On Thu, Oct 28, 2021 at 10:07 PM Frank Rowand <[email protected]> wrote:
>
> On 10/28/21 9:07 PM, Rob Herring wrote:
> > On Thu, Oct 28, 2021 at 7:58 PM <[email protected]> wrote:
> >>
> >> From: Frank Rowand <[email protected]>
> >>
> >> Fix kernel build warning:
> >> drivers/of/unittest-data/tests-interrupts.dtsi:32.26-35.6: Warning (interrupt_map): /testcase-data/interrupts/intmap1: Missing '#address-cells' in interrupt-map provider
> >>
> >> A recently implemented dtc compiler warning reported the dts problem.
> >>
> >> Signed-off-by: Frank Rowand <[email protected]>
> >> ---
> >> drivers/of/unittest-data/tests-interrupts.dtsi | 1 +
> >> 1 file changed, 1 insertion(+)
> >>
> >> diff --git a/drivers/of/unittest-data/tests-interrupts.dtsi b/drivers/of/unittest-data/tests-interrupts.dtsi
> >> index 9b60a549f502..8c2b91b998aa 100644
> >> --- a/drivers/of/unittest-data/tests-interrupts.dtsi
> >> +++ b/drivers/of/unittest-data/tests-interrupts.dtsi
> >> @@ -31,6 +31,7 @@ test_intmap0: intmap0 {
> >>
> >> test_intmap1: intmap1 {
> >> #interrupt-cells = <2>;
> >> + #address-cells = <1>;
> >
> > Notice that we have 2 nodes with interrupt-map here. One has
> > '#address-cells' and one doesn't. Why? Because we need to test that
> > the code can handle both cases.>
> > The dtc warnings are more what should 'new' users do. I don't know
> > what DTs don't have #address-cells, but my guess is ancient ones.
> >
> > Rob
> >
>
> I had hoped to build all of the .dts files in the Linux tree, with the
> new dtc, but did not get to that today. That should flush out any
> cases that would result in build fail from the new approach of treating
> all warnings as errors. I may get to that tomorrow.
They are still just warnings. You mean the requirement to be warning
free? That's not new for dts files.
> If there any any existing .dts files that will trigger the interrupt
> map warning, will we require that they be fixed so that the build will
> not fail?
I already submitted patches for them.
Rob
On 10/29/21 8:16 AM, Rob Herring wrote:
> On Thu, Oct 28, 2021 at 10:07 PM Frank Rowand <[email protected]> wrote:
>>
>> On 10/28/21 9:07 PM, Rob Herring wrote:
>>> On Thu, Oct 28, 2021 at 7:58 PM <[email protected]> wrote:
>>>>
>>>> From: Frank Rowand <[email protected]>
>>>>
>>>> Fix kernel build warning:
>>>> drivers/of/unittest-data/tests-interrupts.dtsi:32.26-35.6: Warning (interrupt_map): /testcase-data/interrupts/intmap1: Missing '#address-cells' in interrupt-map provider
>>>>
>>>> A recently implemented dtc compiler warning reported the dts problem.
>>>>
>>>> Signed-off-by: Frank Rowand <[email protected]>
>>>> ---
>>>> drivers/of/unittest-data/tests-interrupts.dtsi | 1 +
>>>> 1 file changed, 1 insertion(+)
>>>>
>>>> diff --git a/drivers/of/unittest-data/tests-interrupts.dtsi b/drivers/of/unittest-data/tests-interrupts.dtsi
>>>> index 9b60a549f502..8c2b91b998aa 100644
>>>> --- a/drivers/of/unittest-data/tests-interrupts.dtsi
>>>> +++ b/drivers/of/unittest-data/tests-interrupts.dtsi
>>>> @@ -31,6 +31,7 @@ test_intmap0: intmap0 {
>>>>
>>>> test_intmap1: intmap1 {
>>>> #interrupt-cells = <2>;
>>>> + #address-cells = <1>;
>>>
>>> Notice that we have 2 nodes with interrupt-map here. One has
>>> '#address-cells' and one doesn't. Why? Because we need to test that
>>> the code can handle both cases.>
>>> The dtc warnings are more what should 'new' users do. I don't know
>>> what DTs don't have #address-cells, but my guess is ancient ones.
>>>
>>> Rob
>>>
>>
>> I had hoped to build all of the .dts files in the Linux tree, with the
>> new dtc, but did not get to that today. That should flush out any
>> cases that would result in build fail from the new approach of treating
>> all warnings as errors. I may get to that tomorrow.
>
> They are still just warnings. You mean the requirement to be warning
> free? That's not new for dts files.
>
>> If there any any existing .dts files that will trigger the interrupt
>> map warning, will we require that they be fixed so that the build will
>> not fail?
>
> I already submitted patches for them.
OK, I'll drop this patchk, and instead create a patch that documents
why the unittest .dts is missing #address-cells.
-Frank
>
> Rob
>