By reading the code, I found these marcos are never
referenced in the code. Just remove them.

Signed-off-by: Ba Jing <bajing@cmss.chinamobile.com>
---
 drivers/perf/arm-ccn.c | 5 -----
 1 file changed, 5 deletions(-)

diff --git a/drivers/perf/arm-ccn.c b/drivers/perf/arm-ccn.c
index 5c66b9278862..ea878b025d32 100644
--- a/drivers/perf/arm-ccn.c
+++ b/drivers/perf/arm-ccn.c
@@ -28,12 +28,7 @@
 
 #define CCN_MN_ERRINT_STATUS		0x0008
 #define CCN_MN_ERRINT_STATUS__INTREQ__DESSERT		0x11
-#define CCN_MN_ERRINT_STATUS__ALL_ERRORS__ENABLE	0x02
-#define CCN_MN_ERRINT_STATUS__ALL_ERRORS__DISABLED	0x20
 #define CCN_MN_ERRINT_STATUS__ALL_ERRORS__DISABLE	0x22
-#define CCN_MN_ERRINT_STATUS__CORRECTED_ERRORS_ENABLE	0x04
-#define CCN_MN_ERRINT_STATUS__CORRECTED_ERRORS_DISABLED	0x40
-#define CCN_MN_ERRINT_STATUS__CORRECTED_ERRORS_DISABLE	0x44
 #define CCN_MN_ERRINT_STATUS__PMU_EVENTS__ENABLE	0x08
 #define CCN_MN_ERRINT_STATUS__PMU_EVENTS__DISABLED	0x80
 #define CCN_MN_ERRINT_STATUS__PMU_EVENTS__DISABLE	0x88
-- 
2.33.0

On Sat, Oct 12, 2024 at 05:16:29PM +0800, Ba Jing wrote:
> By reading the code, I found these marcos are never
> referenced in the code. Just remove them.
> 
> Signed-off-by: Ba Jing <bajing@cmss.chinamobile.com>

These are control bit values, so I don't see a strong reason to delete
them unless we're certain we're not going to use them in future.

Anyhow, Robin Murphy wrote this driver, so its up to him. Please Cc him
for any changes to this driver in future.

Mark.

> ---
>  drivers/perf/arm-ccn.c | 5 -----
>  1 file changed, 5 deletions(-)
> 
> diff --git a/drivers/perf/arm-ccn.c b/drivers/perf/arm-ccn.c
> index 5c66b9278862..ea878b025d32 100644
> --- a/drivers/perf/arm-ccn.c
> +++ b/drivers/perf/arm-ccn.c
> @@ -28,12 +28,7 @@
>  
>  #define CCN_MN_ERRINT_STATUS		0x0008
>  #define CCN_MN_ERRINT_STATUS__INTREQ__DESSERT		0x11
> -#define CCN_MN_ERRINT_STATUS__ALL_ERRORS__ENABLE	0x02
> -#define CCN_MN_ERRINT_STATUS__ALL_ERRORS__DISABLED	0x20
>  #define CCN_MN_ERRINT_STATUS__ALL_ERRORS__DISABLE	0x22
> -#define CCN_MN_ERRINT_STATUS__CORRECTED_ERRORS_ENABLE	0x04
> -#define CCN_MN_ERRINT_STATUS__CORRECTED_ERRORS_DISABLED	0x40
> -#define CCN_MN_ERRINT_STATUS__CORRECTED_ERRORS_DISABLE	0x44
>  #define CCN_MN_ERRINT_STATUS__PMU_EVENTS__ENABLE	0x08
>  #define CCN_MN_ERRINT_STATUS__PMU_EVENTS__DISABLED	0x80
>  #define CCN_MN_ERRINT_STATUS__PMU_EVENTS__DISABLE	0x88
> -- 
> 2.33.0
> 
> 
>

On 14/10/2024 10:19 am, Mark Rutland wrote:
> On Sat, Oct 12, 2024 at 05:16:29PM +0800, Ba Jing wrote:
>> By reading the code, I found these marcos are never
>> referenced in the code. Just remove them.
>>
>> Signed-off-by: Ba Jing <bajing@cmss.chinamobile.com>
> 
> These are control bit values, so I don't see a strong reason to delete
> them unless we're certain we're not going to use them in future.
> 
> Anyhow, Robin Murphy wrote this driver, so its up to him. Please Cc him
> for any changes to this driver in future.

Er, this is CCN - FWIW I didn't claim this one, nor CCI, as supported in 
my maintainers entry largely because active development on them has long 
finished, but also because I *didn't* write them, so wouldn't claim to 
understand them much better than you or Will do ;)

However, I'd agree that there doesn't seem to be any significant benefit 
to removing these (certainly showing at a glance that ALL_ERRORS_DISABLE 
is composed equivalently to PMU_EVENTS_DISABLE seems like no bad thing.)

Thanks,
Robin.

> 
> Mark.
> 
>> ---
>>   drivers/perf/arm-ccn.c | 5 -----
>>   1 file changed, 5 deletions(-)
>>
>> diff --git a/drivers/perf/arm-ccn.c b/drivers/perf/arm-ccn.c
>> index 5c66b9278862..ea878b025d32 100644
>> --- a/drivers/perf/arm-ccn.c
>> +++ b/drivers/perf/arm-ccn.c
>> @@ -28,12 +28,7 @@
>>   
>>   #define CCN_MN_ERRINT_STATUS		0x0008
>>   #define CCN_MN_ERRINT_STATUS__INTREQ__DESSERT		0x11
>> -#define CCN_MN_ERRINT_STATUS__ALL_ERRORS__ENABLE	0x02
>> -#define CCN_MN_ERRINT_STATUS__ALL_ERRORS__DISABLED	0x20
>>   #define CCN_MN_ERRINT_STATUS__ALL_ERRORS__DISABLE	0x22
>> -#define CCN_MN_ERRINT_STATUS__CORRECTED_ERRORS_ENABLE	0x04
>> -#define CCN_MN_ERRINT_STATUS__CORRECTED_ERRORS_DISABLED	0x40
>> -#define CCN_MN_ERRINT_STATUS__CORRECTED_ERRORS_DISABLE	0x44
>>   #define CCN_MN_ERRINT_STATUS__PMU_EVENTS__ENABLE	0x08
>>   #define CCN_MN_ERRINT_STATUS__PMU_EVENTS__DISABLED	0x80
>>   #define CCN_MN_ERRINT_STATUS__PMU_EVENTS__DISABLE	0x88
>> -- 
>> 2.33.0
>>
>>
>>

On Mon, Oct 14, 2024 at 04:40:52PM +0100, Robin Murphy wrote:
> On 14/10/2024 10:19 am, Mark Rutland wrote:
> > On Sat, Oct 12, 2024 at 05:16:29PM +0800, Ba Jing wrote:
> > > By reading the code, I found these marcos are never
> > > referenced in the code. Just remove them.
> > > 
> > > Signed-off-by: Ba Jing <bajing@cmss.chinamobile.com>
> > 
> > These are control bit values, so I don't see a strong reason to delete
> > them unless we're certain we're not going to use them in future.
> > 
> > Anyhow, Robin Murphy wrote this driver, so its up to him. Please Cc him
> > for any changes to this driver in future.
> 
> Er, this is CCN - FWIW I didn't claim this one, nor CCI, as supported in my
> maintainers entry largely because active development on them has long
> finished, but also because I *didn't* write them, so wouldn't claim to
> understand them much better than you or Will do ;)
> 
> However, I'd agree that there doesn't seem to be any significant benefit to
> removing these (certainly showing at a glance that ALL_ERRORS_DISABLE is
> composed equivalently to PMU_EVENTS_DISABLE seems like no bad thing.)

Yup. FWIW, I wasn't intending to take any of the patches removing unused
#defines in the PMU drivers. I'll pick the one removing an unused
'list_head', though, as that at least makes a structure a bit smaller.

Will