2024-03-01 10:26:44

by Fabrice Gasnier

[permalink] [raw]
Subject: [PATCH v5] counter: Introduce the COUNTER_COMP_FREQUENCY() macro

Now that there are two users for the "frequency" extension, introduce a
new COUNTER_COMP_FREQUENCY() macro.
This extension is intended to be a read-only signal attribute.

Suggested-by: William Breathitt Gray <[email protected]>
Signed-off-by: Fabrice Gasnier <[email protected]>
---
Changes in v5
- "frequency" extension is read-only, so there's no need to provide
a write parameter.
- patch sent separately from "counter: Add stm32 timer events support" [1]
[1] https://lore.kernel.org/lkml/[email protected]/
---
include/linux/counter.h | 7 +++++++
1 file changed, 7 insertions(+)

diff --git a/include/linux/counter.h b/include/linux/counter.h
index 702e9108bbb4..0ac36f815b7d 100644
--- a/include/linux/counter.h
+++ b/include/linux/counter.h
@@ -602,6 +602,13 @@ struct counter_array {
#define COUNTER_COMP_FLOOR(_read, _write) \
COUNTER_COMP_COUNT_U64("floor", _read, _write)

+#define COUNTER_COMP_FREQUENCY(_read) \
+{ \
+ .type = COUNTER_COMP_U64, \
+ .name = "frequency", \
+ .signal_u64_read = (_read), \
+}
+
#define COUNTER_COMP_POLARITY(_read, _write, _available) \
{ \
.type = COUNTER_COMP_SIGNAL_POLARITY, \
--
2.25.1



2024-03-01 16:06:44

by William Breathitt Gray

[permalink] [raw]
Subject: Re: [PATCH v5] counter: Introduce the COUNTER_COMP_FREQUENCY() macro

On Fri, Mar 01, 2024 at 11:25:05AM +0100, Fabrice Gasnier wrote:
> Now that there are two users for the "frequency" extension, introduce a
> new COUNTER_COMP_FREQUENCY() macro.
> This extension is intended to be a read-only signal attribute.
>
> Suggested-by: William Breathitt Gray <[email protected]>
> Signed-off-by: Fabrice Gasnier <[email protected]>
> ---
> Changes in v5
> - "frequency" extension is read-only, so there's no need to provide
> a write parameter.
> - patch sent separately from "counter: Add stm32 timer events support" [1]
> [1] https://lore.kernel.org/lkml/[email protected]/
> ---
> include/linux/counter.h | 7 +++++++
> 1 file changed, 7 insertions(+)
>
> diff --git a/include/linux/counter.h b/include/linux/counter.h
> index 702e9108bbb4..0ac36f815b7d 100644
> --- a/include/linux/counter.h
> +++ b/include/linux/counter.h
> @@ -602,6 +602,13 @@ struct counter_array {
> #define COUNTER_COMP_FLOOR(_read, _write) \
> COUNTER_COMP_COUNT_U64("floor", _read, _write)
>
> +#define COUNTER_COMP_FREQUENCY(_read) \
> +{ \
> + .type = COUNTER_COMP_U64, \
> + .name = "frequency", \
> + .signal_u64_read = (_read), \
> +}
> +
> #define COUNTER_COMP_POLARITY(_read, _write, _available) \
> { \
> .type = COUNTER_COMP_SIGNAL_POLARITY, \
> --
> 2.25.1

Hi Fabrice,

Setting the structure members directly works, but why not use
COUNTER_COMP_SIGNAL_U64("frequency", _read, NULL) instead to keep the
code more succinct?

William Breathitt Gray


Attachments:
(No filename) (1.53 kB)
signature.asc (235.00 B)
Download all attachments

2024-03-04 09:09:37

by Fabrice Gasnier

[permalink] [raw]
Subject: Re: [PATCH v5] counter: Introduce the COUNTER_COMP_FREQUENCY() macro

On 3/1/24 16:55, William Breathitt Gray wrote:
> On Fri, Mar 01, 2024 at 11:25:05AM +0100, Fabrice Gasnier wrote:
>> Now that there are two users for the "frequency" extension, introduce a
>> new COUNTER_COMP_FREQUENCY() macro.
>> This extension is intended to be a read-only signal attribute.
>>
>> Suggested-by: William Breathitt Gray <[email protected]>
>> Signed-off-by: Fabrice Gasnier <[email protected]>
>> ---
>> Changes in v5
>> - "frequency" extension is read-only, so there's no need to provide
>> a write parameter.
>> - patch sent separately from "counter: Add stm32 timer events support" [1]
>> [1] https://lore.kernel.org/lkml/[email protected]/
>> ---
>> include/linux/counter.h | 7 +++++++
>> 1 file changed, 7 insertions(+)
>>
>> diff --git a/include/linux/counter.h b/include/linux/counter.h
>> index 702e9108bbb4..0ac36f815b7d 100644
>> --- a/include/linux/counter.h
>> +++ b/include/linux/counter.h
>> @@ -602,6 +602,13 @@ struct counter_array {
>> #define COUNTER_COMP_FLOOR(_read, _write) \
>> COUNTER_COMP_COUNT_U64("floor", _read, _write)
>>
>> +#define COUNTER_COMP_FREQUENCY(_read) \
>> +{ \
>> + .type = COUNTER_COMP_U64, \
>> + .name = "frequency", \
>> + .signal_u64_read = (_read), \
>> +}
>> +
>> #define COUNTER_COMP_POLARITY(_read, _write, _available) \
>> { \
>> .type = COUNTER_COMP_SIGNAL_POLARITY, \
>> --
>> 2.25.1
>
> Hi Fabrice,
>
> Setting the structure members directly works, but why not use
> COUNTER_COMP_SIGNAL_U64("frequency", _read, NULL) instead to keep the
> code more succinct?

Hi William,

I originally wrote it this way, but I had a doubt since some macros use
the structure members directly.

I can update to use COUNTER_COMP_SIGNAL_U64() instead, that will spare
few lines.

Please let me know what you prefer (I guess your proposal above ?).

Best Regards,
Thanks,
Fabrice

>
> William Breathitt Gray

2024-03-04 13:28:32

by William Breathitt Gray

[permalink] [raw]
Subject: Re: [PATCH v5] counter: Introduce the COUNTER_COMP_FREQUENCY() macro

On Mon, Mar 04, 2024 at 09:41:14AM +0100, Fabrice Gasnier wrote:
> On 3/1/24 16:55, William Breathitt Gray wrote:
> > On Fri, Mar 01, 2024 at 11:25:05AM +0100, Fabrice Gasnier wrote:
> >> Now that there are two users for the "frequency" extension, introduce a
> >> new COUNTER_COMP_FREQUENCY() macro.
> >> This extension is intended to be a read-only signal attribute.
> >>
> >> Suggested-by: William Breathitt Gray <[email protected]>
> >> Signed-off-by: Fabrice Gasnier <[email protected]>
> >> ---
> >> Changes in v5
> >> - "frequency" extension is read-only, so there's no need to provide
> >> a write parameter.
> >> - patch sent separately from "counter: Add stm32 timer events support" [1]
> >> [1] https://lore.kernel.org/lkml/[email protected]/
> >> ---
> >> include/linux/counter.h | 7 +++++++
> >> 1 file changed, 7 insertions(+)
> >>
> >> diff --git a/include/linux/counter.h b/include/linux/counter.h
> >> index 702e9108bbb4..0ac36f815b7d 100644
> >> --- a/include/linux/counter.h
> >> +++ b/include/linux/counter.h
> >> @@ -602,6 +602,13 @@ struct counter_array {
> >> #define COUNTER_COMP_FLOOR(_read, _write) \
> >> COUNTER_COMP_COUNT_U64("floor", _read, _write)
> >>
> >> +#define COUNTER_COMP_FREQUENCY(_read) \
> >> +{ \
> >> + .type = COUNTER_COMP_U64, \
> >> + .name = "frequency", \
> >> + .signal_u64_read = (_read), \
> >> +}
> >> +
> >> #define COUNTER_COMP_POLARITY(_read, _write, _available) \
> >> { \
> >> .type = COUNTER_COMP_SIGNAL_POLARITY, \
> >> --
> >> 2.25.1
> >
> > Hi Fabrice,
> >
> > Setting the structure members directly works, but why not use
> > COUNTER_COMP_SIGNAL_U64("frequency", _read, NULL) instead to keep the
> > code more succinct?
>
> Hi William,
>
> I originally wrote it this way, but I had a doubt since some macros use
> the structure members directly.

Ah yes, the macros that use the members directly are typically the ones
that are unique for their particular type. For example, the enum
constant type COUNTER_COMP_COUNT_DIRECTION will only ever be used with
the COUNTER_COMP_DIRECTION() macro.

For macros that are based on general types such as COUNTER_COMP_U64,
it's better to use the respective base macro such as
COUNTER_COMP_SIGNAL_U64(). Not only is this more succinct and clearer of
the intent, if the need arises in the future it allows us to upgrade the
the underlying base macro and have those changes propagate to the macros
that utilize it.

>
> I can update to use COUNTER_COMP_SIGNAL_U64() instead, that will spare
> few lines.
>
> Please let me know what you prefer (I guess your proposal above ?).
>
> Best Regards,
> Thanks,
> Fabrice

Update to use COUNTER_COMP_SIGNAL_U64() instead, and I should be able to
pick it up quickly.

Thanks,

William Breathitt Gray


Attachments:
(No filename) (2.82 kB)
signature.asc (235.00 B)
Download all attachments