* [arm-perf:for-next/perf 6/8] drivers/perf/arm-ni.c:433:13: warning: operator '?:' has lower precedence than '+'; '+' will be evaluated first
@ 2025-11-04 9:43 kernel test robot
2025-11-04 12:35 ` Will Deacon
0 siblings, 1 reply; 7+ messages in thread
From: kernel test robot @ 2025-11-04 9:43 UTC (permalink / raw)
To: Robin Murphy; +Cc: llvm, oe-kbuild-all, Will Deacon
tree: https://git.kernel.org/pub/scm/linux/kernel/git/will/linux.git for-next/perf
head: 989b40b7578a2be7b0388522d33d751b257d59d9
commit: 8fa08f8835e57e1fd5e2994c9cf76c84dadc1235 [6/8] perf/arm-ni: Add NoC S3 support
config: sparc64-randconfig-002-20251104 (https://download.01.org/0day-ci/archive/20251104/202511041749.ok8zDP6u-lkp@intel.com/config)
compiler: clang version 22.0.0git (https://github.com/llvm/llvm-project d2625a438020ad35330cda29c3def102c1687b1b)
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20251104/202511041749.ok8zDP6u-lkp@intel.com/reproduce)
If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Closes: https://lore.kernel.org/oe-kbuild-all/202511041749.ok8zDP6u-lkp@intel.com/
All warnings (new ones prefixed by >>):
>> drivers/perf/arm-ni.c:433:13: warning: operator '?:' has lower precedence than '+'; '+' will be evaluated first [-Wparentheses]
432 | hw->event_base = (unsigned long)cd->pmu_base +
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
433 | is_7xx ? NI700_PMCCNTR_L : NI_PMCCNTR_L;
| ~~~~~~ ^
drivers/perf/arm-ni.c:433:13: note: place parentheses around the '+' expression to silence this warning
432 | hw->event_base = (unsigned long)cd->pmu_base +
| (
433 | is_7xx ? NI700_PMCCNTR_L : NI_PMCCNTR_L;
| ^
| )
drivers/perf/arm-ni.c:433:13: note: place parentheses around the '?:' expression to evaluate it first
433 | is_7xx ? NI700_PMCCNTR_L : NI_PMCCNTR_L;
| ~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/perf/arm-ni.c:446:13: warning: operator '?:' has lower precedence than '+'; '+' will be evaluated first [-Wparentheses]
445 | hw->event_base = (unsigned long)cd->pmu_base +
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
446 | is_7xx ? NI700_PMEVCNTR(hw->idx) : NI_PMEVCNTR(hw->idx);
| ~~~~~~ ^
drivers/perf/arm-ni.c:446:13: note: place parentheses around the '+' expression to silence this warning
445 | hw->event_base = (unsigned long)cd->pmu_base +
| (
446 | is_7xx ? NI700_PMEVCNTR(hw->idx) : NI_PMEVCNTR(hw->idx);
| ^
| )
drivers/perf/arm-ni.c:446:13: note: place parentheses around the '?:' expression to evaluate it first
446 | is_7xx ? NI700_PMEVCNTR(hw->idx) : NI_PMEVCNTR(hw->idx);
| ~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2 warnings generated.
vim +433 drivers/perf/arm-ni.c
418
419 static int arm_ni_event_add(struct perf_event *event, int flags)
420 {
421 struct arm_ni_cd *cd = pmu_to_cd(event->pmu);
422 struct hw_perf_event *hw = &event->hw;
423 struct arm_ni_unit *unit;
424 enum ni_node_type type = NI_EVENT_TYPE(event);
425 bool is_7xx = arm_ni_is_7xx(cd_to_ni(cd));
426 u32 reg;
427
428 if (type == NI_PMU) {
429 if (cd->ccnt)
430 return -ENOSPC;
431 hw->idx = NI_CCNT_IDX;
432 hw->event_base = (unsigned long)cd->pmu_base +
> 433 is_7xx ? NI700_PMCCNTR_L : NI_PMCCNTR_L;
434 cd->ccnt = event;
435 arm_ni_init_ccnt(hw);
436 } else {
437 hw->idx = 0;
438 while (cd->evcnt[hw->idx]) {
439 if (++hw->idx == NI_NUM_COUNTERS)
440 return -ENOSPC;
441 }
442 cd->evcnt[hw->idx] = event;
443 unit = (void *)hw->config_base;
444 unit->event[hw->idx] = NI_EVENT_EVENTID(event);
445 hw->event_base = (unsigned long)cd->pmu_base +
446 is_7xx ? NI700_PMEVCNTR(hw->idx) : NI_PMEVCNTR(hw->idx);
447 arm_ni_init_evcnt(hw);
448 lo_hi_writeq_relaxed(le64_to_cpu(unit->pmusel), unit->pmusela);
449
450 reg = FIELD_PREP(NI_PMEVTYPER_NODE_TYPE, type) |
451 FIELD_PREP(NI_PMEVTYPER_NODE_ID, NI_EVENT_NODEID(event));
452 writel_relaxed(reg, cd->pmu_base + NI_PMEVTYPER(hw->idx));
453 }
454 if (flags & PERF_EF_START)
455 arm_ni_event_start(event, 0);
456 return 0;
457 }
458
--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [arm-perf:for-next/perf 6/8] drivers/perf/arm-ni.c:433:13: warning: operator '?:' has lower precedence than '+'; '+' will be evaluated first
2025-11-04 9:43 [arm-perf:for-next/perf 6/8] drivers/perf/arm-ni.c:433:13: warning: operator '?:' has lower precedence than '+'; '+' will be evaluated first kernel test robot
@ 2025-11-04 12:35 ` Will Deacon
2025-11-04 13:26 ` Robin Murphy
0 siblings, 1 reply; 7+ messages in thread
From: Will Deacon @ 2025-11-04 12:35 UTC (permalink / raw)
To: kernel test robot; +Cc: Robin Murphy, llvm, oe-kbuild-all
On Tue, Nov 04, 2025 at 05:43:40PM +0800, kernel test robot wrote:
> tree: https://git.kernel.org/pub/scm/linux/kernel/git/will/linux.git for-next/perf
> head: 989b40b7578a2be7b0388522d33d751b257d59d9
> commit: 8fa08f8835e57e1fd5e2994c9cf76c84dadc1235 [6/8] perf/arm-ni: Add NoC S3 support
> config: sparc64-randconfig-002-20251104 (https://download.01.org/0day-ci/archive/20251104/202511041749.ok8zDP6u-lkp@intel.com/config)
> compiler: clang version 22.0.0git (https://github.com/llvm/llvm-project d2625a438020ad35330cda29c3def102c1687b1b)
> reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20251104/202511041749.ok8zDP6u-lkp@intel.com/reproduce)
>
> If you fix the issue in a separate patch/commit (i.e. not just a new version of
> the same patch/commit), kindly add following tags
> | Reported-by: kernel test robot <lkp@intel.com>
> | Closes: https://lore.kernel.org/oe-kbuild-all/202511041749.ok8zDP6u-lkp@intel.com/
>
> All warnings (new ones prefixed by >>):
>
> >> drivers/perf/arm-ni.c:433:13: warning: operator '?:' has lower precedence than '+'; '+' will be evaluated first [-Wparentheses]
> 432 | hw->event_base = (unsigned long)cd->pmu_base +
> | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> 433 | is_7xx ? NI700_PMCCNTR_L : NI_PMCCNTR_L;
> | ~~~~~~ ^
> drivers/perf/arm-ni.c:433:13: note: place parentheses around the '+' expression to silence this warning
> 432 | hw->event_base = (unsigned long)cd->pmu_base +
> | (
> 433 | is_7xx ? NI700_PMCCNTR_L : NI_PMCCNTR_L;
> | ^
> | )
Hrm. Doesn't that mean this is completely broken or am I misreading the
report?
Will
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [arm-perf:for-next/perf 6/8] drivers/perf/arm-ni.c:433:13: warning: operator '?:' has lower precedence than '+'; '+' will be evaluated first
2025-11-04 12:35 ` Will Deacon
@ 2025-11-04 13:26 ` Robin Murphy
2025-11-04 13:32 ` Will Deacon
0 siblings, 1 reply; 7+ messages in thread
From: Robin Murphy @ 2025-11-04 13:26 UTC (permalink / raw)
To: Will Deacon, kernel test robot; +Cc: llvm, oe-kbuild-all, nd
On 2025-11-04 12:35 pm, Will Deacon wrote:
> On Tue, Nov 04, 2025 at 05:43:40PM +0800, kernel test robot wrote:
>> tree: https://git.kernel.org/pub/scm/linux/kernel/git/will/linux.git for-next/perf
>> head: 989b40b7578a2be7b0388522d33d751b257d59d9
>> commit: 8fa08f8835e57e1fd5e2994c9cf76c84dadc1235 [6/8] perf/arm-ni: Add NoC S3 support
>> config: sparc64-randconfig-002-20251104 (https://download.01.org/0day-ci/archive/20251104/202511041749.ok8zDP6u-lkp@intel.com/config)
>> compiler: clang version 22.0.0git (https://github.com/llvm/llvm-project d2625a438020ad35330cda29c3def102c1687b1b)
>> reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20251104/202511041749.ok8zDP6u-lkp@intel.com/reproduce)
>>
>> If you fix the issue in a separate patch/commit (i.e. not just a new version of
>> the same patch/commit), kindly add following tags
>> | Reported-by: kernel test robot <lkp@intel.com>
>> | Closes: https://lore.kernel.org/oe-kbuild-all/202511041749.ok8zDP6u-lkp@intel.com/
>>
>> All warnings (new ones prefixed by >>):
>>
>>>> drivers/perf/arm-ni.c:433:13: warning: operator '?:' has lower precedence than '+'; '+' will be evaluated first [-Wparentheses]
>> 432 | hw->event_base = (unsigned long)cd->pmu_base +
>> | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>> 433 | is_7xx ? NI700_PMCCNTR_L : NI_PMCCNTR_L;
>> | ~~~~~~ ^
>> drivers/perf/arm-ni.c:433:13: note: place parentheses around the '+' expression to silence this warning
>> 432 | hw->event_base = (unsigned long)cd->pmu_base +
>> | (
>> 433 | is_7xx ? NI700_PMCCNTR_L : NI_PMCCNTR_L;
>> | ^
>> | )
>
> Hrm. Doesn't that mean this is completely broken or am I misreading the
> report?
Bah, my arm64 W=1 build (which I did do in the hope of catching any such
sillies...) fails to flag that :(
Truth is this is a fresh refactoring of the mess I originally wrote that
got tested by someone else 18 months ago, and I don't have any hardware
to functionally test it myself. Feel free to drop it and I'll send a
complete v2, or I can do an incremental fix if you prefer (I already
realised I could have streamlined the version check to a one-off in
event_init as well.)
Cheers,
Robin.
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [arm-perf:for-next/perf 6/8] drivers/perf/arm-ni.c:433:13: warning: operator '?:' has lower precedence than '+'; '+' will be evaluated first
2025-11-04 13:26 ` Robin Murphy
@ 2025-11-04 13:32 ` Will Deacon
2025-11-04 14:04 ` Robin Murphy
0 siblings, 1 reply; 7+ messages in thread
From: Will Deacon @ 2025-11-04 13:32 UTC (permalink / raw)
To: Robin Murphy; +Cc: kernel test robot, llvm, oe-kbuild-all, nd
On Tue, Nov 04, 2025 at 01:26:11PM +0000, Robin Murphy wrote:
> On 2025-11-04 12:35 pm, Will Deacon wrote:
> > On Tue, Nov 04, 2025 at 05:43:40PM +0800, kernel test robot wrote:
> > > tree: https://git.kernel.org/pub/scm/linux/kernel/git/will/linux.git for-next/perf
> > > head: 989b40b7578a2be7b0388522d33d751b257d59d9
> > > commit: 8fa08f8835e57e1fd5e2994c9cf76c84dadc1235 [6/8] perf/arm-ni: Add NoC S3 support
> > > config: sparc64-randconfig-002-20251104 (https://download.01.org/0day-ci/archive/20251104/202511041749.ok8zDP6u-lkp@intel.com/config)
> > > compiler: clang version 22.0.0git (https://github.com/llvm/llvm-project d2625a438020ad35330cda29c3def102c1687b1b)
> > > reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20251104/202511041749.ok8zDP6u-lkp@intel.com/reproduce)
> > >
> > > If you fix the issue in a separate patch/commit (i.e. not just a new version of
> > > the same patch/commit), kindly add following tags
> > > | Reported-by: kernel test robot <lkp@intel.com>
> > > | Closes: https://lore.kernel.org/oe-kbuild-all/202511041749.ok8zDP6u-lkp@intel.com/
> > >
> > > All warnings (new ones prefixed by >>):
> > >
> > > > > drivers/perf/arm-ni.c:433:13: warning: operator '?:' has lower precedence than '+'; '+' will be evaluated first [-Wparentheses]
> > > 432 | hw->event_base = (unsigned long)cd->pmu_base +
> > > | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > > 433 | is_7xx ? NI700_PMCCNTR_L : NI_PMCCNTR_L;
> > > | ~~~~~~ ^
> > > drivers/perf/arm-ni.c:433:13: note: place parentheses around the '+' expression to silence this warning
> > > 432 | hw->event_base = (unsigned long)cd->pmu_base +
> > > | (
> > > 433 | is_7xx ? NI700_PMCCNTR_L : NI_PMCCNTR_L;
> > > | ^
> > > | )
> >
> > Hrm. Doesn't that mean this is completely broken or am I misreading the
> > report?
>
> Bah, my arm64 W=1 build (which I did do in the hope of catching any such
> sillies...) fails to flag that :(
>
> Truth is this is a fresh refactoring of the mess I originally wrote that got
> tested by someone else 18 months ago, and I don't have any hardware to
> functionally test it myself. Feel free to drop it and I'll send a complete
> v2, or I can do an incremental fix if you prefer (I already realised I could
> have streamlined the version check to a one-off in event_init as well.)
Damn, it's not ideal that we don't have any test coverage for this.
The patch is buried under some others now, so I can either revert it or
take fixes on top. Up to you.
Will
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [arm-perf:for-next/perf 6/8] drivers/perf/arm-ni.c:433:13: warning: operator '?:' has lower precedence than '+'; '+' will be evaluated first
2025-11-04 13:32 ` Will Deacon
@ 2025-11-04 14:04 ` Robin Murphy
2025-11-04 18:24 ` Will Deacon
0 siblings, 1 reply; 7+ messages in thread
From: Robin Murphy @ 2025-11-04 14:04 UTC (permalink / raw)
To: Will Deacon; +Cc: kernel test robot, llvm, oe-kbuild-all, nd
On 2025-11-04 1:32 pm, Will Deacon wrote:
> On Tue, Nov 04, 2025 at 01:26:11PM +0000, Robin Murphy wrote:
>> On 2025-11-04 12:35 pm, Will Deacon wrote:
>>> On Tue, Nov 04, 2025 at 05:43:40PM +0800, kernel test robot wrote:
>>>> tree: https://git.kernel.org/pub/scm/linux/kernel/git/will/linux.git for-next/perf
>>>> head: 989b40b7578a2be7b0388522d33d751b257d59d9
>>>> commit: 8fa08f8835e57e1fd5e2994c9cf76c84dadc1235 [6/8] perf/arm-ni: Add NoC S3 support
>>>> config: sparc64-randconfig-002-20251104 (https://download.01.org/0day-ci/archive/20251104/202511041749.ok8zDP6u-lkp@intel.com/config)
>>>> compiler: clang version 22.0.0git (https://github.com/llvm/llvm-project d2625a438020ad35330cda29c3def102c1687b1b)
>>>> reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20251104/202511041749.ok8zDP6u-lkp@intel.com/reproduce)
>>>>
>>>> If you fix the issue in a separate patch/commit (i.e. not just a new version of
>>>> the same patch/commit), kindly add following tags
>>>> | Reported-by: kernel test robot <lkp@intel.com>
>>>> | Closes: https://lore.kernel.org/oe-kbuild-all/202511041749.ok8zDP6u-lkp@intel.com/
>>>>
>>>> All warnings (new ones prefixed by >>):
>>>>
>>>>>> drivers/perf/arm-ni.c:433:13: warning: operator '?:' has lower precedence than '+'; '+' will be evaluated first [-Wparentheses]
>>>> 432 | hw->event_base = (unsigned long)cd->pmu_base +
>>>> | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>>>> 433 | is_7xx ? NI700_PMCCNTR_L : NI_PMCCNTR_L;
>>>> | ~~~~~~ ^
>>>> drivers/perf/arm-ni.c:433:13: note: place parentheses around the '+' expression to silence this warning
>>>> 432 | hw->event_base = (unsigned long)cd->pmu_base +
>>>> | (
>>>> 433 | is_7xx ? NI700_PMCCNTR_L : NI_PMCCNTR_L;
>>>> | ^
>>>> | )
>>>
>>> Hrm. Doesn't that mean this is completely broken or am I misreading the
>>> report?
>>
>> Bah, my arm64 W=1 build (which I did do in the hope of catching any such
>> sillies...) fails to flag that :(
>>
>> Truth is this is a fresh refactoring of the mess I originally wrote that got
>> tested by someone else 18 months ago, and I don't have any hardware to
>> functionally test it myself. Feel free to drop it and I'll send a complete
>> v2, or I can do an incremental fix if you prefer (I already realised I could
>> have streamlined the version check to a one-off in event_init as well.)
>
> Damn, it's not ideal that we don't have any test coverage for this.
>
> The patch is buried under some others now, so I can either revert it or
> take fixes on top. Up to you.
I've managed to break my git-send-email setup at the moment, so I hope
an inline patch is OK!
Cheers,
Robin.
----->8-----
Subject: [PATCH] perf/arm-ni: Fix and optimise register offset calculation
LKP points out an operator precedence oversight in the new NoC S3
support that, annoyingly, my local W=1 build didn't flag. In fixing
that, we can also take the similarly-missed opportunity to cache the
version check itself at event_init time.
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202511041749.ok8zDP6u-lkp@intel.com/
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
---
drivers/perf/arm-ni.c | 21 ++++++++++++---------
1 file changed, 12 insertions(+), 9 deletions(-)
diff --git a/drivers/perf/arm-ni.c b/drivers/perf/arm-ni.c
index aa824abc629e..66858c65215d 100644
--- a/drivers/perf/arm-ni.c
+++ b/drivers/perf/arm-ni.c
@@ -323,9 +323,15 @@ static int arm_ni_validate_group(struct perf_event *event)
return 0;
}
+static bool arm_ni_is_7xx(const struct arm_ni *ni)
+{
+ return ni->part == PART_NI_700 || ni->part == PART_NI_710AE;
+}
+
static int arm_ni_event_init(struct perf_event *event)
{
struct arm_ni_cd *cd = pmu_to_cd(event->pmu);
+ struct arm_ni *ni;
if (event->attr.type != event->pmu->type)
return -ENOENT;
@@ -333,7 +339,10 @@ static int arm_ni_event_init(struct perf_event *event)
if (is_sampling_event(event))
return -EINVAL;
- event->cpu = cd_to_ni(cd)->cpu;
+ ni = cd_to_ni(cd);
+ event->cpu = ni->cpu;
+ event->hw.flags = arm_ni_is_7xx(ni);
+
if (NI_EVENT_TYPE(event) == NI_PMU)
return arm_ni_validate_group(event);
@@ -411,18 +420,12 @@ static void arm_ni_init_evcnt(struct hw_perf_event *hw)
writel_relaxed(S32_MIN, (void __iomem *)hw->event_base);
}
-static bool arm_ni_is_7xx(const struct arm_ni *ni)
-{
- return ni->part == PART_NI_700 || ni->part == PART_NI_710AE;
-}
-
static int arm_ni_event_add(struct perf_event *event, int flags)
{
struct arm_ni_cd *cd = pmu_to_cd(event->pmu);
struct hw_perf_event *hw = &event->hw;
struct arm_ni_unit *unit;
enum ni_node_type type = NI_EVENT_TYPE(event);
- bool is_7xx = arm_ni_is_7xx(cd_to_ni(cd));
u32 reg;
if (type == NI_PMU) {
@@ -430,7 +433,7 @@ static int arm_ni_event_add(struct perf_event *event, int flags)
return -ENOSPC;
hw->idx = NI_CCNT_IDX;
hw->event_base = (unsigned long)cd->pmu_base +
- is_7xx ? NI700_PMCCNTR_L : NI_PMCCNTR_L;
+ (hw->flags ? NI700_PMCCNTR_L : NI_PMCCNTR_L);
cd->ccnt = event;
arm_ni_init_ccnt(hw);
} else {
@@ -443,7 +446,7 @@ static int arm_ni_event_add(struct perf_event *event, int flags)
unit = (void *)hw->config_base;
unit->event[hw->idx] = NI_EVENT_EVENTID(event);
hw->event_base = (unsigned long)cd->pmu_base +
- is_7xx ? NI700_PMEVCNTR(hw->idx) : NI_PMEVCNTR(hw->idx);
+ (hw->flags ? NI700_PMEVCNTR(hw->idx) : NI_PMEVCNTR(hw->idx));
arm_ni_init_evcnt(hw);
lo_hi_writeq_relaxed(le64_to_cpu(unit->pmusel), unit->pmusela);
--
2.34.1
^ permalink raw reply related [flat|nested] 7+ messages in thread
* Re: [arm-perf:for-next/perf 6/8] drivers/perf/arm-ni.c:433:13: warning: operator '?:' has lower precedence than '+'; '+' will be evaluated first
2025-11-04 14:04 ` Robin Murphy
@ 2025-11-04 18:24 ` Will Deacon
2025-11-04 18:36 ` Robin Murphy
0 siblings, 1 reply; 7+ messages in thread
From: Will Deacon @ 2025-11-04 18:24 UTC (permalink / raw)
To: Robin Murphy; +Cc: kernel test robot, llvm, oe-kbuild-all, nd
On Tue, Nov 04, 2025 at 02:04:57PM +0000, Robin Murphy wrote:
> On 2025-11-04 1:32 pm, Will Deacon wrote:
> > On Tue, Nov 04, 2025 at 01:26:11PM +0000, Robin Murphy wrote:
> > > On 2025-11-04 12:35 pm, Will Deacon wrote:
> > > > On Tue, Nov 04, 2025 at 05:43:40PM +0800, kernel test robot wrote:
> > > > > tree: https://git.kernel.org/pub/scm/linux/kernel/git/will/linux.git for-next/perf
> > > > > head: 989b40b7578a2be7b0388522d33d751b257d59d9
> > > > > commit: 8fa08f8835e57e1fd5e2994c9cf76c84dadc1235 [6/8] perf/arm-ni: Add NoC S3 support
> > > > > config: sparc64-randconfig-002-20251104 (https://download.01.org/0day-ci/archive/20251104/202511041749.ok8zDP6u-lkp@intel.com/config)
> > > > > compiler: clang version 22.0.0git (https://github.com/llvm/llvm-project d2625a438020ad35330cda29c3def102c1687b1b)
> > > > > reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20251104/202511041749.ok8zDP6u-lkp@intel.com/reproduce)
> > > > >
> > > > > If you fix the issue in a separate patch/commit (i.e. not just a new version of
> > > > > the same patch/commit), kindly add following tags
> > > > > | Reported-by: kernel test robot <lkp@intel.com>
> > > > > | Closes: https://lore.kernel.org/oe-kbuild-all/202511041749.ok8zDP6u-lkp@intel.com/
> > > > >
> > > > > All warnings (new ones prefixed by >>):
> > > > >
> > > > > > > drivers/perf/arm-ni.c:433:13: warning: operator '?:' has lower precedence than '+'; '+' will be evaluated first [-Wparentheses]
> > > > > 432 | hw->event_base = (unsigned long)cd->pmu_base +
> > > > > | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > > > > 433 | is_7xx ? NI700_PMCCNTR_L : NI_PMCCNTR_L;
> > > > > | ~~~~~~ ^
> > > > > drivers/perf/arm-ni.c:433:13: note: place parentheses around the '+' expression to silence this warning
> > > > > 432 | hw->event_base = (unsigned long)cd->pmu_base +
> > > > > | (
> > > > > 433 | is_7xx ? NI700_PMCCNTR_L : NI_PMCCNTR_L;
> > > > > | ^
> > > > > | )
> > > >
> > > > Hrm. Doesn't that mean this is completely broken or am I misreading the
> > > > report?
> > >
> > > Bah, my arm64 W=1 build (which I did do in the hope of catching any such
> > > sillies...) fails to flag that :(
> > >
> > > Truth is this is a fresh refactoring of the mess I originally wrote that got
> > > tested by someone else 18 months ago, and I don't have any hardware to
> > > functionally test it myself. Feel free to drop it and I'll send a complete
> > > v2, or I can do an incremental fix if you prefer (I already realised I could
> > > have streamlined the version check to a one-off in event_init as well.)
> >
> > Damn, it's not ideal that we don't have any test coverage for this.
> >
> > The patch is buried under some others now, so I can either revert it or
> > take fixes on top. Up to you.
>
> I've managed to break my git-send-email setup at the moment, so I hope
> an inline patch is OK!
Sadly, this appears to be whitespace corrupted in a very strange way. The
version I see in mutt is different to the version on lore, which itself
renders differently when viewed as "raw". None of them apply :(
Will
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [arm-perf:for-next/perf 6/8] drivers/perf/arm-ni.c:433:13: warning: operator '?:' has lower precedence than '+'; '+' will be evaluated first
2025-11-04 18:24 ` Will Deacon
@ 2025-11-04 18:36 ` Robin Murphy
0 siblings, 0 replies; 7+ messages in thread
From: Robin Murphy @ 2025-11-04 18:36 UTC (permalink / raw)
To: Will Deacon; +Cc: kernel test robot, llvm, oe-kbuild-all
On 2025-11-04 6:24 pm, Will Deacon wrote:
> On Tue, Nov 04, 2025 at 02:04:57PM +0000, Robin Murphy wrote:
>> On 2025-11-04 1:32 pm, Will Deacon wrote:
>>> On Tue, Nov 04, 2025 at 01:26:11PM +0000, Robin Murphy wrote:
>>>> On 2025-11-04 12:35 pm, Will Deacon wrote:
>>>>> On Tue, Nov 04, 2025 at 05:43:40PM +0800, kernel test robot wrote:
>>>>>> tree: https://git.kernel.org/pub/scm/linux/kernel/git/will/linux.git for-next/perf
>>>>>> head: 989b40b7578a2be7b0388522d33d751b257d59d9
>>>>>> commit: 8fa08f8835e57e1fd5e2994c9cf76c84dadc1235 [6/8] perf/arm-ni: Add NoC S3 support
>>>>>> config: sparc64-randconfig-002-20251104 (https://download.01.org/0day-ci/archive/20251104/202511041749.ok8zDP6u-lkp@intel.com/config)
>>>>>> compiler: clang version 22.0.0git (https://github.com/llvm/llvm-project d2625a438020ad35330cda29c3def102c1687b1b)
>>>>>> reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20251104/202511041749.ok8zDP6u-lkp@intel.com/reproduce)
>>>>>>
>>>>>> If you fix the issue in a separate patch/commit (i.e. not just a new version of
>>>>>> the same patch/commit), kindly add following tags
>>>>>> | Reported-by: kernel test robot <lkp@intel.com>
>>>>>> | Closes: https://lore.kernel.org/oe-kbuild-all/202511041749.ok8zDP6u-lkp@intel.com/
>>>>>>
>>>>>> All warnings (new ones prefixed by >>):
>>>>>>
>>>>>>>> drivers/perf/arm-ni.c:433:13: warning: operator '?:' has lower precedence than '+'; '+' will be evaluated first [-Wparentheses]
>>>>>> 432 | hw->event_base = (unsigned long)cd->pmu_base +
>>>>>> | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>>>>>> 433 | is_7xx ? NI700_PMCCNTR_L : NI_PMCCNTR_L;
>>>>>> | ~~~~~~ ^
>>>>>> drivers/perf/arm-ni.c:433:13: note: place parentheses around the '+' expression to silence this warning
>>>>>> 432 | hw->event_base = (unsigned long)cd->pmu_base +
>>>>>> | (
>>>>>> 433 | is_7xx ? NI700_PMCCNTR_L : NI_PMCCNTR_L;
>>>>>> | ^
>>>>>> | )
>>>>>
>>>>> Hrm. Doesn't that mean this is completely broken or am I misreading the
>>>>> report?
>>>>
>>>> Bah, my arm64 W=1 build (which I did do in the hope of catching any such
>>>> sillies...) fails to flag that :(
>>>>
>>>> Truth is this is a fresh refactoring of the mess I originally wrote that got
>>>> tested by someone else 18 months ago, and I don't have any hardware to
>>>> functionally test it myself. Feel free to drop it and I'll send a complete
>>>> v2, or I can do an incremental fix if you prefer (I already realised I could
>>>> have streamlined the version check to a one-off in event_init as well.)
>>>
>>> Damn, it's not ideal that we don't have any test coverage for this.
>>>
>>> The patch is buried under some others now, so I can either revert it or
>>> take fixes on top. Up to you.
>>
>> I've managed to break my git-send-email setup at the moment, so I hope
>> an inline patch is OK!
>
> Sadly, this appears to be whitespace corrupted in a very strange way. The
> version I see in mutt is different to the version on lore, which itself
> renders differently when viewed as "raw". None of them apply :(
Hooray for Office 365 and its liberal approximation of IMAP... Anyway, I
have now managed to sort out the machine that was keeping me out of my
other account, so you should have a properly-sent version now.
Cheers,
Robin.
^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2025-11-04 18:36 UTC | newest]
Thread overview: 7+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-11-04 9:43 [arm-perf:for-next/perf 6/8] drivers/perf/arm-ni.c:433:13: warning: operator '?:' has lower precedence than '+'; '+' will be evaluated first kernel test robot
2025-11-04 12:35 ` Will Deacon
2025-11-04 13:26 ` Robin Murphy
2025-11-04 13:32 ` Will Deacon
2025-11-04 14:04 ` Robin Murphy
2025-11-04 18:24 ` Will Deacon
2025-11-04 18:36 ` Robin Murphy
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox