Re: [PATCH V3 0/7] iommu/vt-d: Support performance monitoring for IOMMU

["Liang, Kan" <kan.liang@linux.intel.com>]

On 2023-01-28 2:10 a.m., Baolu Lu wrote:
> On 2023/1/21 0:54, kan.liang@linux.intel.com wrote:
>> From: Kan Liang <kan.liang@linux.intel.com>
>>
>> Changes since V2:
>> - Move ecmd_submit_sync() to iommu.c to avoid #ifdef CONFIG_INTEL_IOMMU
>>
>> Changes since V1:
>> - The cap and ecap registers are always in the first page. It's not
>>    necessary to use the reg size in dmar_validate_one_drhd(). (Patch 1)
>> - Move reg_size up and pair it with reg_base_addr in struct
>>    dmar_drhd_unit (Patch 1)
>> - Update the year of Copyright (Patch 2)
>> - Return 0 if PMS is not supported (Patch 2)
>> - Refine the comments and add a pr_warn for per-counter capabilities
>>    check (Patch 2)
>> - Remove unnecessary iommu_pmu->num_cntr = i (Patch 2)
>> - Remove has_ob of ecmd_submit_sync() (Patch 3)
>> - Remove the helpers from non-INTEL_IOMMU. (Patch 3)
>> - Still keep #ifdef CONFIG_INTEL_IOMMU for ecmd_submit_sync() to
>>    avoid compile warning with non-INTEL_IOMMU config.
>> - Use pr_warn_once() to replace WARN_ONCE() (Patch 4 & 6)
>> - Free iommu PMU if it fails to be registered. (Patch 4)
>> - Remove unnecessary 'handled' variable. (Patch 6)
>>
>> A performance monitoring infrastructure, perfmon, is introduced with
>> the VT-d Spec 4.0. The purpose of perfmon is to support collection of
>> information about key events occurring during operation of the remapping
>> hardware, to aid performance tuning and debug. The patch series is to
>> support the perfmon for IOMMU.
>>
>> To facilitate the perfmon support, the patch series also supports two
>> new generic features of VT-d Spec 4.0.
>> - Support the 'size' field to retrieve the accurate size of the register
>>    set for each dmar device from DRHD. (Patch 1)
>> - Support the new Enhanced Command Interface. (Patch 3)
>>
>> With the patch series, users can collect the performance data of IOMMU
>> via Linux perf tool. For example,
>>
>>   $ perf stat -e dmar0/iommu_requests,filter_ats=0/ -a sleep 1
>>
>>   Performance counter stats for 'system wide':
>>
>>                2135      dmar0/iommu_requests,filter_ats=0/
>>
>>         1.001087695 seconds time elapsed
>>
>> The IOMMU PMUs can be found under /sys/bus/event_source/devices/dmar*
>>
>> The available filters and event format can be found at the format folder
>>   $ ls /sys/bus/event_source/devices/dmar0/format/
>> event  event_group  filter_ats  filter_page_table
>>
>> The supported events can be found at the events folder
>>
>>   $ ls /sys/bus/event_source/devices/dmar0/events/
>> ats_blocked        int_cache_hit_nonposted  iommu_mrds
>> pasid_cache_lookup
>> ctxt_cache_hit     int_cache_hit_posted     iommu_requests
>> pg_req_posted
>> ctxt_cache_lookup  int_cache_lookup         iotlb_hit
>> pw_occupancy
>> fs_nonleaf_hit     iommu_clocks             iotlb_lookup
>> ss_nonleaf_hit
>> fs_nonleaf_lookup  iommu_mem_blocked        pasid_cache_hit
>> ss_nonleaf_lookup
>>
>> Kan Liang (7):
>>    iommu/vt-d: Support size of the register set in DRHD
>>    iommu/vt-d: Retrieve IOMMU perfmon capability information
>>    iommu/vt-d: Support Enhanced Command Interface
>>    iommu/vt-d: Add IOMMU perfmon support
>>    iommu/vt-d: Support cpumask for IOMMU perfmon
>>    iommu/vt-d: Add IOMMU perfmon overflow handler support
>>    iommu/vt-d: Enable IOMMU perfmon support
>>
>>   .../sysfs-bus-event_source-devices-iommu      |  32 +
>>   drivers/iommu/intel/Kconfig                   |   9 +
>>   drivers/iommu/intel/Makefile                  |   1 +
>>   drivers/iommu/intel/dmar.c                    |  33 +-
>>   drivers/iommu/intel/iommu.c                   |  59 ++
>>   drivers/iommu/intel/iommu.h                   | 101 +-
>>   drivers/iommu/intel/perfmon.c                 | 860 ++++++++++++++++++
>>   drivers/iommu/intel/perfmon.h                 |  65 ++
>>   drivers/iommu/intel/svm.c                     |   2 +-
>>   include/acpi/actbl1.h                         |   2 +-
>>   include/linux/cpuhotplug.h                    |   1 +
>>   include/linux/dmar.h                          |   1 +
>>   12 files changed, 1159 insertions(+), 7 deletions(-)
>>   create mode 100644
>> Documentation/ABI/testing/sysfs-bus-event_source-devices-iommu
>>   create mode 100644 drivers/iommu/intel/perfmon.c
>>   create mode 100644 drivers/iommu/intel/perfmon.h
> 
> Thanks for the work. The overall looks good to me now. But I am still
> seeing some code style issues after running "./scripts/checkpatch.pl
> --strict" scripts.
> 
> Can you please fix those issues and post a v4?
> 
Sure.

Thanks,
Kan