Re: [PATCH 1/2] arm: perf: Fix ARCH=arm build with GCC in armv8pmu_write_evtype()

From: James Clark <james.clark@arm.com>
To: Mark Rutland <mark.rutland@arm.com>
Cc: linux-arm-kernel@lists.infradead.org,
	linux-perf-users@vger.kernel.org, linux-next@vger.kernel.org,
	will@kernel.org, u.kleine-koenig@pengutronix.de,
	Nathan Chancellor <nathan@kernel.org>,
	Nick Desaulniers <ndesaulniers@google.com>,
	Bill Wendling <morbo@google.com>,
	Justin Stitt <justinstitt@google.com>,
	Anshuman Khandual <anshuman.khandual@arm.com>,
	Suzuki K Poulose <suzuki.poulose@arm.com>,
	linux-kernel@vger.kernel.org, llvm@lists.linux.dev
Subject: Re: [PATCH 1/2] arm: perf: Fix ARCH=arm build with GCC in armv8pmu_write_evtype()
Date: Fri, 15 Dec 2023 17:42:50 +0000	[thread overview]
Message-ID: <d335905e-38bb-1994-57f6-533852b7a560@arm.com> (raw)
In-Reply-To: <ZXx9GW1onSy4eBEB@FVFF77S0Q05N>

On 15/12/2023 16:21, Mark Rutland wrote:
> On Fri, Dec 15, 2023 at 03:00:38PM +0000, James Clark wrote:
>> LLVM ignores everything inside the if statement and doesn't generate
>> errors, but GCC doesn't ignore it, resulting in the following error:
>>
>>   drivers/perf/arm_pmuv3.c: In function 'armv8pmu_write_evtype':
>>   include/linux/bits.h:34:29: error: left shift count >= width of type [-Werror=shift-count-overflow]
>>   	34 |         (((~UL(0)) - (UL(1) << (l)) + 1) & \
>>
>> Fix it by changing the if to #if.
> 
> I reckon it'd be cleaner to use GENMASK_ULL for the TH and TC fields, in
> include/linux/perf/arm_pmu.h have:
> 
> | /*
> |  * PMXEVTYPER: Event selection reg
> |  */
> | #define ARMV8_PMU_EVTYPE_EVENT GENMASK(15, 0)          /* Mask for EVENT bits */
> | #define ARMV8_PMU_EVTYPE_TH    GENMASK_ULL(43, 32)     /* arm64 only */
> | #define ARMV8_PMU_EVTYPE_TC    GENMASK_ULL(63, 61)     /* arm64 only */
> 
> IIUC that should silence this warning, and it'd remove the need for the
> ifdeffery and other changes in patch 2.
> 
> Does that work, or am I missing something?
> 
> Thanks,
> Mark.
> 

You're right that does work. For some reason I thought there was some
component of writing it to the mask that was the issue as well.

I'll send another version with that fix instead.

Thanks
James

>>
>> Fixes: 3115ee021bfb ("arm64: perf: Include threshold control fields in PMEVTYPER mask")
>> Reported-by: Uwe Kleine-K"onig <u.kleine-koenig@pengutronix.de>
>> Closes: https://lore.kernel.org/linux-arm-kernel/20231215120817.h2f3akgv72zhrtqo@pengutronix.de/
>> Signed-off-by: James Clark <james.clark@arm.com>
>> ---
>>  drivers/perf/arm_pmuv3.c | 5 +++--
>>  1 file changed, 3 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/perf/arm_pmuv3.c b/drivers/perf/arm_pmuv3.c
>> index 23fa6c5da82c..3ed2086cefc3 100644
>> --- a/drivers/perf/arm_pmuv3.c
>> +++ b/drivers/perf/arm_pmuv3.c
>> @@ -631,8 +631,9 @@ static void armv8pmu_write_evtype(int idx, unsigned long val)
>>  			     ARMV8_PMU_EXCLUDE_EL0 |
>>  			     ARMV8_PMU_EXCLUDE_EL1;
>>  
>> -	if (IS_ENABLED(CONFIG_ARM64))
>> -		mask |= ARMV8_PMU_EVTYPE_TC | ARMV8_PMU_EVTYPE_TH;
>> +#if IS_ENABLED(CONFIG_ARM64)
>> +	mask |= ARMV8_PMU_EVTYPE_TC | ARMV8_PMU_EVTYPE_TH;
>> +#endif
>>  
>>  	val &= mask;
>>  	write_pmevtypern(counter, val);
>> -- 
>> 2.34.1
>>