drm/amdgpu: add pmu counters
authorJonathan Kim <jonathan.kim@amd.com>
Thu, 20 Jun 2019 03:56:25 +0000 (23:56 -0400)
committerAlex Deucher <alexander.deucher@amd.com>
Thu, 20 Jun 2019 16:36:22 +0000 (11:36 -0500)
adding perf event counters

Signed-off-by: Jonathan Kim <Jonathan.Kim@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drivers/gpu/drm/amd/amdgpu/Makefile
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
drivers/gpu/drm/amd/amdgpu/amdgpu_pmu.c [new file with mode: 0644]
drivers/gpu/drm/amd/amdgpu/amdgpu_pmu.h [new file with mode: 0644]

index 57ce44cc32268fed9e660365062840f9391effd1..4c9fd2645f64c241d2f98de810e65e1b2f06752d 100644 (file)
@@ -54,7 +54,7 @@ amdgpu-y += amdgpu_device.o amdgpu_kms.o \
        amdgpu_gtt_mgr.o amdgpu_vram_mgr.o amdgpu_virt.o amdgpu_atomfirmware.o \
        amdgpu_vf_error.o amdgpu_sched.o amdgpu_debugfs.o amdgpu_ids.o \
        amdgpu_gmc.o amdgpu_xgmi.o amdgpu_csa.o amdgpu_ras.o amdgpu_vm_cpu.o \
-       amdgpu_vm_sdma.o
+       amdgpu_vm_sdma.o amdgpu_pmu.o
 
 # add asic specific block
 amdgpu-$(CONFIG_DRM_AMDGPU_CIK)+= cik.o cik_ih.o kv_smc.o kv_dpm.o \
index e26303bc567e5b8146d3f0569fd6bcce2afff351..e6ddd30f31a6ef46a4213d3ef10a4930e619123d 100644 (file)
@@ -61,6 +61,7 @@
 
 #include "amdgpu_xgmi.h"
 #include "amdgpu_ras.h"
+#include "amdgpu_pmu.h"
 
 MODULE_FIRMWARE("amdgpu/vega10_gpu_info.bin");
 MODULE_FIRMWARE("amdgpu/vega12_gpu_info.bin");
@@ -2746,6 +2747,10 @@ fence_driver_init:
                return r;
        }
 
+       r = amdgpu_pmu_init(adev);
+       if (r)
+               dev_err(adev->dev, "amdgpu_pmu_init failed\n");
+
        return 0;
 
 failed:
@@ -2814,6 +2819,7 @@ void amdgpu_device_fini(struct amdgpu_device *adev)
        amdgpu_debugfs_regs_cleanup(adev);
        device_remove_file(adev->dev, &dev_attr_pcie_replay_count);
        amdgpu_ucode_sysfs_fini(adev);
+       amdgpu_pmu_fini(adev);
 }
 
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_pmu.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_pmu.c
new file mode 100644 (file)
index 0000000..0e6dba9
--- /dev/null
@@ -0,0 +1,280 @@
+/*
+ * Copyright 2019 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Author: Jonathan Kim <jonathan.kim@amd.com>
+ *
+ */
+
+#include <linux/perf_event.h>
+#include <linux/init.h>
+#include "amdgpu.h"
+#include "amdgpu_pmu.h"
+#include "df_v3_6.h"
+
+#define PMU_NAME_SIZE 32
+
+/* record to keep track of pmu entry per pmu type per device */
+struct amdgpu_pmu_entry {
+       struct list_head entry;
+       struct amdgpu_device *adev;
+       struct pmu pmu;
+       unsigned int pmu_perf_type;
+};
+
+static LIST_HEAD(amdgpu_pmu_list);
+
+
+/* initialize perf counter */
+static int amdgpu_perf_event_init(struct perf_event *event)
+{
+       struct hw_perf_event *hwc = &event->hw;
+
+       /* test the event attr type check for PMU enumeration */
+       if (event->attr.type != event->pmu->type)
+               return -ENOENT;
+
+       /* update the hw_perf_event struct with config data */
+       hwc->conf = event->attr.config;
+
+       return 0;
+}
+
+/* start perf counter */
+static void amdgpu_perf_start(struct perf_event *event, int flags)
+{
+       struct hw_perf_event *hwc = &event->hw;
+       struct amdgpu_pmu_entry *pe = container_of(event->pmu,
+                                                 struct amdgpu_pmu_entry,
+                                                 pmu);
+
+       if (WARN_ON_ONCE(!(hwc->state & PERF_HES_STOPPED)))
+               return;
+
+       WARN_ON_ONCE(!(hwc->state & PERF_HES_UPTODATE));
+       hwc->state = 0;
+
+       switch (pe->pmu_perf_type) {
+       case PERF_TYPE_AMDGPU_DF:
+               if (!(flags & PERF_EF_RELOAD))
+                       pe->adev->df_funcs->pmc_start(pe->adev, hwc->conf, 1);
+
+               pe->adev->df_funcs->pmc_start(pe->adev, hwc->conf, 0);
+               break;
+       default:
+               break;
+       }
+
+       perf_event_update_userpage(event);
+
+}
+
+/* read perf counter */
+static void amdgpu_perf_read(struct perf_event *event)
+{
+       struct hw_perf_event *hwc = &event->hw;
+       struct amdgpu_pmu_entry *pe = container_of(event->pmu,
+                                                 struct amdgpu_pmu_entry,
+                                                 pmu);
+
+       u64 count, prev;
+
+       do {
+               prev = local64_read(&hwc->prev_count);
+
+               switch (pe->pmu_perf_type) {
+               case PERF_TYPE_AMDGPU_DF:
+                       pe->adev->df_funcs->pmc_get_count(pe->adev, hwc->conf,
+                                                         &count);
+                       break;
+               default:
+                       count = 0;
+                       break;
+               };
+       } while (local64_cmpxchg(&hwc->prev_count, prev, count) != prev);
+
+       local64_add(count - prev, &event->count);
+}
+
+/* stop perf counter */
+static void amdgpu_perf_stop(struct perf_event *event, int flags)
+{
+       struct hw_perf_event *hwc = &event->hw;
+       struct amdgpu_pmu_entry *pe = container_of(event->pmu,
+                                                 struct amdgpu_pmu_entry,
+                                                 pmu);
+
+       if (hwc->state & PERF_HES_UPTODATE)
+               return;
+
+       switch (pe->pmu_perf_type) {
+       case PERF_TYPE_AMDGPU_DF:
+               pe->adev->df_funcs->pmc_stop(pe->adev, hwc->conf, 0);
+               break;
+       default:
+               break;
+       };
+
+       WARN_ON_ONCE(hwc->state & PERF_HES_STOPPED);
+       hwc->state |= PERF_HES_STOPPED;
+
+       if (hwc->state & PERF_HES_UPTODATE)
+               return;
+
+       amdgpu_perf_read(event);
+       hwc->state |= PERF_HES_UPTODATE;
+}
+
+/* add perf counter  */
+static int amdgpu_perf_add(struct perf_event *event, int flags)
+{
+       struct hw_perf_event *hwc = &event->hw;
+       int retval;
+
+       struct amdgpu_pmu_entry *pe = container_of(event->pmu,
+                                                 struct amdgpu_pmu_entry,
+                                                 pmu);
+
+       event->hw.state = PERF_HES_UPTODATE | PERF_HES_STOPPED;
+
+       switch (pe->pmu_perf_type) {
+       case PERF_TYPE_AMDGPU_DF:
+               retval = pe->adev->df_funcs->pmc_start(pe->adev, hwc->conf, 1);
+               break;
+       default:
+               return 0;
+       };
+
+       if (retval)
+               return retval;
+
+       if (flags & PERF_EF_START)
+               amdgpu_perf_start(event, PERF_EF_RELOAD);
+
+       return retval;
+
+}
+
+/* delete perf counter  */
+static void amdgpu_perf_del(struct perf_event *event, int flags)
+{
+       struct hw_perf_event *hwc = &event->hw;
+       struct amdgpu_pmu_entry *pe = container_of(event->pmu,
+                                                 struct amdgpu_pmu_entry,
+                                                 pmu);
+
+       amdgpu_perf_stop(event, PERF_EF_UPDATE);
+
+       switch (pe->pmu_perf_type) {
+       case PERF_TYPE_AMDGPU_DF:
+               pe->adev->df_funcs->pmc_stop(pe->adev, hwc->conf, 1);
+               break;
+       default:
+               break;
+       };
+
+       perf_event_update_userpage(event);
+}
+
+/* vega20 pmus */
+
+/* init pmu tracking per pmu type */
+static int init_pmu_by_type(struct amdgpu_device *adev,
+                 const struct attribute_group *attr_groups[],
+                 char *pmu_type_name, char *pmu_file_prefix,
+                 unsigned int pmu_perf_type,
+                 unsigned int num_counters)
+{
+       char pmu_name[PMU_NAME_SIZE];
+       struct amdgpu_pmu_entry *pmu_entry;
+       int ret = 0;
+
+       pmu_entry = kzalloc(sizeof(struct amdgpu_pmu_entry), GFP_KERNEL);
+
+       if (!pmu_entry)
+               return -ENOMEM;
+
+       pmu_entry->adev = adev;
+       pmu_entry->pmu = (struct pmu){
+               .event_init = amdgpu_perf_event_init,
+               .add = amdgpu_perf_add,
+               .del = amdgpu_perf_del,
+               .start = amdgpu_perf_start,
+               .stop = amdgpu_perf_stop,
+               .read = amdgpu_perf_read,
+               .task_ctx_nr = perf_invalid_context,
+       };
+
+       pmu_entry->pmu.attr_groups = attr_groups;
+       pmu_entry->pmu_perf_type = pmu_perf_type;
+       snprintf(pmu_name, PMU_NAME_SIZE, "%s_%d",
+                               pmu_file_prefix, adev->ddev->primary->index);
+
+       ret = perf_pmu_register(&pmu_entry->pmu, pmu_name, -1);
+
+       if (ret) {
+               kfree(pmu_entry);
+               pr_warn("Error initializing AMDGPU %s PMUs.\n", pmu_type_name);
+               return ret;
+       }
+
+       pr_info("Detected AMDGPU %s Counters. # of Counters = %d.\n",
+                       pmu_type_name, num_counters);
+
+       list_add_tail(&pmu_entry->entry, &amdgpu_pmu_list);
+
+       return 0;
+}
+
+/* init amdgpu_pmu */
+int amdgpu_pmu_init(struct amdgpu_device *adev)
+{
+       int ret = 0;
+
+       switch (adev->asic_type) {
+       case CHIP_VEGA20:
+               /* init df */
+               ret = init_pmu_by_type(adev, df_v3_6_attr_groups,
+                                      "DF", "amdgpu_df", PERF_TYPE_AMDGPU_DF,
+                                      DF_V3_6_MAX_COUNTERS);
+
+               /* other pmu types go here*/
+               break;
+       default:
+               return 0;
+       }
+
+       return 0;
+}
+
+
+/* destroy all pmu data associated with target device */
+void amdgpu_pmu_fini(struct amdgpu_device *adev)
+{
+       struct amdgpu_pmu_entry *pe, *temp;
+
+       list_for_each_entry_safe(pe, temp, &amdgpu_pmu_list, entry) {
+               if (pe->adev == adev) {
+                       list_del(&pe->entry);
+                       perf_pmu_unregister(&pe->pmu);
+                       kfree(pe);
+               }
+       }
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_pmu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_pmu.h
new file mode 100644 (file)
index 0000000..7dddb71
--- /dev/null
@@ -0,0 +1,37 @@
+/*
+ * Copyright 2019 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Author: Jonathan Kim <jonathan.kim@amd.com>
+ *
+ */
+
+#ifndef _AMDGPU_PMU_H_
+#define _AMDGPU_PMU_H_
+
+enum amdgpu_pmu_perf_type {
+       PERF_TYPE_AMDGPU_DF = 0,
+       PERF_TYPE_AMDGPU_MAX
+};
+
+int amdgpu_pmu_init(struct amdgpu_device *adev);
+void amdgpu_pmu_fini(struct amdgpu_device *adev);
+
+#endif /* _AMDGPU_PMU_H_ */