drm/amd/amdgpu: Add tracepoint for DMA page mapping (v4)

This helps map DMA addresses back to physical addresses.

Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

(v2):  Added tracepoints for USERPTR, SG mappings, and
     SWIOTBL mappings.  Reformatted trace call perform
     PCI decoding internal to the trace.

(v3):  Add unmap tracepoints as well

(v4):  Move traces into separate functions

2 files changed, 106 insertions, 6 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h
index 9ab58245e518..d8cd3e554488 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h
@@ -14,6 +14,62 @@
 #define AMDGPU_JOB_GET_TIMELINE_NAME(job) \
          job->base.s_fence->finished.ops->get_timeline_name(&job->base.s_fence->finished)
 
+TRACE_EVENT(amdgpu_ttm_tt_populate,
+            TP_PROTO(struct amdgpu_device *adev, uint64_t dma_address, uint64_t phys_address),
+            TP_ARGS(adev, dma_address, phys_address),
+            TP_STRUCT__entry(
+                                __field(uint16_t, domain)
+                                __field(uint8_t, bus)
+                                __field(uint8_t, slot)
+                                __field(uint8_t, func)
+                                __field(uint64_t, dma)
+                                __field(uint64_t, phys)
+                            ),
+            TP_fast_assign(
+                           __entry->domain = pci_domain_nr(adev->pdev->bus);
+                           __entry->bus = adev->pdev->bus->number;
+                           __entry->slot = PCI_SLOT(adev->pdev->devfn);
+                           __entry->func = PCI_FUNC(adev->pdev->devfn);
+                           __entry->dma = dma_address;
+                           __entry->phys = phys_address;
+                           ),
+            TP_printk("%04x:%02x:%02x.%x: 0x%llx => 0x%llx",
+                      (unsigned)__entry->domain,
+                      (unsigned)__entry->bus,
+                      (unsigned)__entry->slot,
+                      (unsigned)__entry->func,
+                      (unsigned long long)__entry->dma,
+                      (unsigned long long)__entry->phys)
+);
+
+TRACE_EVENT(amdgpu_ttm_tt_unpopulate,
+            TP_PROTO(struct amdgpu_device *adev, uint64_t dma_address, uint64_t phys_address),
+            TP_ARGS(adev, dma_address, phys_address),
+            TP_STRUCT__entry(
+                                __field(uint16_t, domain)
+                                __field(uint8_t, bus)
+                                __field(uint8_t, slot)
+                                __field(uint8_t, func)
+                                __field(uint64_t, dma)
+                                __field(uint64_t, phys)
+                            ),
+            TP_fast_assign(
+                           __entry->domain = pci_domain_nr(adev->pdev->bus);
+                           __entry->bus = adev->pdev->bus->number;
+                           __entry->slot = PCI_SLOT(adev->pdev->devfn);
+                           __entry->func = PCI_FUNC(adev->pdev->devfn);
+                           __entry->dma = dma_address;
+                           __entry->phys = phys_address;
+                           ),
+            TP_printk("%04x:%02x:%02x.%x: 0x%llx => 0x%llx",
+                      (unsigned)__entry->domain,
+                      (unsigned)__entry->bus,
+                      (unsigned)__entry->slot,
+                      (unsigned)__entry->func,
+                      (unsigned long long)__entry->dma,
+                      (unsigned long long)__entry->phys)
+);
+
 TRACE_EVENT(amdgpu_mm_rreg,
             TP_PROTO(unsigned did, uint32_t reg, uint32_t value),
             TP_ARGS(did, reg, value),
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index c803b082324d..00817b9712df 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -43,6 +43,7 @@
 #include <linux/pagemap.h>
 #include <linux/debugfs.h>
 #include "amdgpu.h"
+#include "amdgpu_trace.h"
 #include "bif/bif_4_1_d.h"
 
 #define DRM_FILE_PAGE_OFFSET (0x100000000ULL >> PAGE_SHIFT)
@@ -662,6 +663,38 @@ release_pages:
         return r;
 }
 
+static void amdgpu_trace_dma_map(struct ttm_tt *ttm)
+{
+        struct amdgpu_device *adev = amdgpu_ttm_adev(ttm->bdev);
+        struct amdgpu_ttm_tt *gtt = (void *)ttm;
+        unsigned i;
+
+        if (unlikely(trace_amdgpu_ttm_tt_populate_enabled())) {
+                for (i = 0; i < ttm->num_pages; i++) {
+                        trace_amdgpu_ttm_tt_populate(
+                                adev,
+                                gtt->ttm.dma_address[i],
+                                page_to_phys(ttm->pages[i]));
+                }
+        }
+}
+
+static void amdgpu_trace_dma_unmap(struct ttm_tt *ttm)
+{
+        struct amdgpu_device *adev = amdgpu_ttm_adev(ttm->bdev);
+        struct amdgpu_ttm_tt *gtt = (void *)ttm;
+        unsigned i;
+
+        if (unlikely(trace_amdgpu_ttm_tt_unpopulate_enabled())) {
+                for (i = 0; i < ttm->num_pages; i++) {
+                        trace_amdgpu_ttm_tt_unpopulate(
+                                adev,
+                                gtt->ttm.dma_address[i],
+                                page_to_phys(ttm->pages[i]));
+                }
+        }
+}
+
 /* prepare the sg table with the user pages */
 static int amdgpu_ttm_tt_pin_userptr(struct ttm_tt *ttm)
 {
@@ -688,6 +721,8 @@ static int amdgpu_ttm_tt_pin_userptr(struct ttm_tt *ttm)
         drm_prime_sg_to_page_addr_arrays(ttm->sg, ttm->pages,
                                          gtt->ttm.dma_address, ttm->num_pages);
 
+        amdgpu_trace_dma_map(ttm);
+
         return 0;
 
 release_sg:
@@ -721,6 +756,8 @@ static void amdgpu_ttm_tt_unpin_userptr(struct ttm_tt *ttm)
                 put_page(page);
         }
 
+        amdgpu_trace_dma_unmap(ttm);
+
         sg_free_table(ttm->sg);
 }
 
@@ -892,7 +929,7 @@ static struct ttm_tt *amdgpu_ttm_tt_create(struct ttm_bo_device *bdev,
 
 static int amdgpu_ttm_tt_populate(struct ttm_tt *ttm)
 {
-        struct amdgpu_device *adev;
+        struct amdgpu_device *adev = amdgpu_ttm_adev(ttm->bdev);
         struct amdgpu_ttm_tt *gtt = (void *)ttm;
         unsigned i;
         int r;
@@ -915,14 +952,14 @@ static int amdgpu_ttm_tt_populate(struct ttm_tt *ttm)
                 drm_prime_sg_to_page_addr_arrays(ttm->sg, ttm->pages,
                                                  gtt->ttm.dma_address, ttm->num_pages);
                 ttm->state = tt_unbound;
-                return 0;
+                r = 0;
+                goto trace_mappings;
         }
 
-        adev = amdgpu_ttm_adev(ttm->bdev);
-
 #ifdef CONFIG_SWIOTLB
         if (swiotlb_nr_tbl()) {
-                return ttm_dma_populate(&gtt->ttm, adev->dev);
+                r = ttm_dma_populate(&gtt->ttm, adev->dev);
+                goto trace_mappings;
         }
 #endif
 
@@ -945,7 +982,12 @@ static int amdgpu_ttm_tt_populate(struct ttm_tt *ttm)
                         return -EFAULT;
                 }
         }
-        return 0;
+
+        r = 0;
+trace_mappings:
+        if (likely(!r))
+                amdgpu_trace_dma_map(ttm);
+        return r;
 }
 
 static void amdgpu_ttm_tt_unpopulate(struct ttm_tt *ttm)
@@ -966,6 +1008,8 @@ static void amdgpu_ttm_tt_unpopulate(struct ttm_tt *ttm)
 
         adev = amdgpu_ttm_adev(ttm->bdev);
 
+        amdgpu_trace_dma_unmap(ttm);
+
 #ifdef CONFIG_SWIOTLB
         if (swiotlb_nr_tbl()) {
                 ttm_dma_unpopulate(&gtt->ttm, adev->dev);