Re: [PATCH V12 05/10] arm64/perf: Add branch stack support in ARMV8 PMU

[Anshuman Khandual]

On 6/19/23 14:38, Marc Zyngier wrote:
> On Mon, 19 Jun 2023 06:45:07 +0100,
> Anshuman Khandual <anshuman.khandual@xxxxxxx> wrote:
>>
>>
>>
>> On 6/16/23 14:51, Catalin Marinas wrote:
>>> On Fri, Jun 16, 2023 at 06:57:52AM +0530, Anshuman Khandual wrote:
>>>> On 6/16/23 05:12, kernel test robot wrote:
>>>>> kernel test robot noticed the following build errors:
>>>>>
>>>>> [auto build test ERROR on arm64/for-next/core]
>>>>> [also build test ERROR on tip/perf/core acme/perf/core linus/master v6.4-rc6 next-20230615]
>>>>> [If your patch is applied to the wrong git tree, kindly drop us a note.
>>>>> And when submitting patch, we suggest to use '--base' as documented in
>>>>> https://git-scm.com/docs/git-format-patch#_base_tree_information]
>>>>>
>>>>> url:    https://github.com/intel-lab-lkp/linux/commits/Anshuman-Khandual/drivers-perf-arm_pmu-Add-new-sched_task-callback/20230615-223352
>>>>> base:   https://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux.git for-next/core
>>>>> patch link:    https://lore.kernel.org/r/20230615133239.442736-6-anshuman.khandual%40arm.com
>>>>> patch subject: [PATCH V12 05/10] arm64/perf: Add branch stack support in ARMV8 PMU
>>>>> config: arm-randconfig-r004-20230615 (https://download.01.org/0day-ci/archive/20230616/202306160706.Uei5XDoi-lkp@xxxxxxxxx/config)
>>>>> compiler: clang version 17.0.0 (https://github.com/llvm/llvm-project.git 4a5ac14ee968ff0ad5d2cc1ffa0299048db4c88a)
>>>>> reproduce (this is a W=1 build):
>>>>>         mkdir -p ~/bin
>>>>>         wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
>>>>>         chmod +x ~/bin/make.cross
>>>>>         # install arm cross compiling tool for clang build
>>>>>         # apt-get install binutils-arm-linux-gnueabi
>>>>>         git remote add arm64 https://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux.git
>>>>>         git fetch arm64 for-next/core
>>>>>         git checkout arm64/for-next/core
>>>>>         b4 shazam https://lore.kernel.org/r/20230615133239.442736-6-anshuman.khandual@xxxxxxx
>>>>>         # save the config file
>>>>>         mkdir build_dir && cp config build_dir/.config
>>>>>         COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang ~/bin/make.cross W=1 O=build_dir ARCH=arm olddefconfig
>>>>>         COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang ~/bin/make.cross W=1 O=build_dir ARCH=arm SHELL=/bin/bash drivers/perf/
>>>>
>>>> I am unable to reproduce this on mainline 6.4-rc6 via default cross compiler
>>>> on a W=1 build. Looking at all other problems reported on the file, it seems
>>>> something is not right here. Reported build problems around these callbacks,
>>>> i.e armv8pmu_branch_XXXX() do not make sense as they are available via config
>>>> CONFIG_PERF_EVENTS which is also enabled along with CONFIG_ARM_PMUV3 in this
>>>> test config.
>>>
>>> Have you tried applying this series on top of the arm64 for-next/core
>>> branch? That's what the robot it testing (in the absence of a --base
>>> option when generating the patches).
>>
>> Right, it turned out to be a build problem on arm (32 bit) platform instead.
>> After arm_pmuv3.c moved into common ./drivers/perf from ./arch/arm64/kernel/,
>> it can no longer access arch/arm64/include/asm/perf_event.h defined functions
>> without breaking arm (32) bit. The following code block needs to be moved out
>> from arch/arm64/include/asm/perf_event.h into include/linux/perf/arm_pmuv3.h
>> (which is preferred as all call sites are inside drivers/perf/arm_pmuv3.c) or
>> may be arm_pmu.h (which is one step higher in the abstraction).
> 
> No, that's the wrong approach. The 32bit backend must have its own
> stubs for the stuff it implements or not.

Okay.


> 
> Just add something like the patch below, and please *test* that a
> 32bit VM using PMUv3 doesn't have any regression.

Sure.

> 
> Thanks,
> 
> 	M.
> 
>>From 017362ca518e6d6ac3262514d1f7f27e73232799 Mon Sep 17 00:00:00 2001
> From: Marc Zyngier <maz@xxxxxxxxxx>
> Date: Mon, 19 Jun 2023 10:05:52 +0100
> Subject: [PATCH] 32bit hack
> 
> Signed-off-by: Marc Zyngier <maz@xxxxxxxxxx>
> ---
>  arch/arm/include/asm/arm_pmuv3.h | 18 ++++++++++++++++++
>  1 file changed, 18 insertions(+)
> 
> diff --git a/arch/arm/include/asm/arm_pmuv3.h b/arch/arm/include/asm/arm_pmuv3.h
> index f4db3e75d75f..c4bcb7a18267 100644
> --- a/arch/arm/include/asm/arm_pmuv3.h
> +++ b/arch/arm/include/asm/arm_pmuv3.h
> @@ -244,4 +244,22 @@ static inline bool is_pmuv3p5(int pmuver)
>  	return pmuver >= ARMV8_PMU_DFR_VER_V3P5;
>  }
>  
> +/* BRBE stubs */

These stubs also need to be wrapped around with #ifdef CONFIG_PERF_EVENTS

> +static inline void armv8pmu_branch_enable(struct perf_event *event) { }
> +static inline void armv8pmu_branch_disable(struct perf_event *event) { }
> +static inline void armv8pmu_branch_read(struct pmu_hw_events * cpuc,
> +					struct perf_event *event) { }
> +static inline void armv8pmu_branch_save(struct arm_pmu *armpmu, void *ctx) {}
> +static inline void armv8pmu_branch_reset(void) {}
> +static inline bool armv8pmu_branch_attr_valid(struct perf_event *event)
> +{
> +	return false;
> +}
> +static inline void armv8pmu_branch_probe(struct arm_pmu *armpmu) {}
> +static inline int armv8pmu_task_ctx_cache_alloc(struct arm_pmu *armpmu)
> +{
> +	return 0;
> +}
> +static inline void armv8pmu_task_ctx_cache_free(struct arm_pmu *armpmu) {}
> +
>  #endif

Sure, will make all the necessary changes.