drm/radeon: Add radeon.test parameter for running BO GPU copy tests.

If enabled, during initialization BO GTT->VRAM and VRAM->GTT GPU copies are
tested across the whole GTT aperture.

This has helped uncover the benchmark copy size bug and verify the maximum
aperture size supported by the AGP bridge in my PowerBook.

Signed-off-by: Michel Dänzer <daenzer@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

5 files changed, 225 insertions, 1 deletions

diff --git a/drivers/gpu/drm/radeon/Makefile b/drivers/gpu/drm/radeon/Makefile
index 5fae1e074b4b..013d38059943 100644
--- a/drivers/gpu/drm/radeon/Makefile
+++ b/drivers/gpu/drm/radeon/Makefile
@@ -13,7 +13,8 @@ radeon-$(CONFIG_DRM_RADEON_KMS) += radeon_device.o radeon_kms.o \
         radeon_encoders.o radeon_display.o radeon_cursor.o radeon_i2c.o \
         radeon_clocks.o radeon_fb.o radeon_gem.o radeon_ring.o radeon_irq_kms.o \
         radeon_cs.o radeon_bios.o radeon_benchmark.o r100.o r300.o r420.o \
-        rs400.o rs600.o rs690.o rv515.o r520.o r600.o rs780.o rv770.o
+        rs400.o rs600.o rs690.o rv515.o r520.o r600.o rs780.o rv770.o \
+        radeon_test.o
 
 radeon-$(CONFIG_COMPAT) += radeon_ioc32.o
 
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index 045b33b3bf2d..b1d945b8ed6c 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -64,6 +64,7 @@ extern int radeon_agpmode;
 extern int radeon_vram_limit;
 extern int radeon_gart_size;
 extern int radeon_benchmarking;
+extern int radeon_testing;
 extern int radeon_connector_table;
 
 /*
@@ -535,6 +536,12 @@ void radeon_benchmark(struct radeon_device *rdev);
 
 
 /*
+ * Testing
+ */
+void radeon_test_moves(struct radeon_device *rdev);
+
+
+/*
  * Debugfs
  */
 int radeon_debugfs_add_files(struct radeon_device *rdev,
diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c
index 6d1749e44222..a162ade74b7f 100644
--- a/drivers/gpu/drm/radeon/radeon_device.c
+++ b/drivers/gpu/drm/radeon/radeon_device.c
@@ -638,6 +638,9 @@ int radeon_device_init(struct radeon_device *rdev,
         if (!ret) {
                 DRM_INFO("radeon: kernel modesetting successfully initialized.\n");
         }
+        if (radeon_testing) {
+                radeon_test_moves(rdev);
+        }
         if (radeon_benchmarking) {
                 radeon_benchmark(rdev);
         }
diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c
index 84ba69f48784..3cfcee17dc56 100644
--- a/drivers/gpu/drm/radeon/radeon_drv.c
+++ b/drivers/gpu/drm/radeon/radeon_drv.c
@@ -89,6 +89,7 @@ int radeon_agpmode = 0;
 int radeon_vram_limit = 0;
 int radeon_gart_size = 512; /* default gart size */
 int radeon_benchmarking = 0;
+int radeon_testing = 0;
 int radeon_connector_table = 0;
 #endif
 
@@ -117,6 +118,9 @@ module_param_named(gartsize, radeon_gart_size, int, 0600);
 MODULE_PARM_DESC(benchmark, "Run benchmark");
 module_param_named(benchmark, radeon_benchmarking, int, 0444);
 
+MODULE_PARM_DESC(test, "Run tests");
+module_param_named(test, radeon_testing, int, 0444);
+
 MODULE_PARM_DESC(connector_table, "Force connector table");
 module_param_named(connector_table, radeon_connector_table, int, 0444);
 #endif
diff --git a/drivers/gpu/drm/radeon/radeon_test.c b/drivers/gpu/drm/radeon/radeon_test.c
new file mode 100644
index 000000000000..03c33cf4e14c
--- /dev/null
+++ b/drivers/gpu/drm/radeon/radeon_test.c
@@ -0,0 +1,209 @@
+/*
+ * Copyright 2009 VMware, 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.
+ *
+ * Authors: Michel Dänzer
+ */
+#include <drm/drmP.h>
+#include <drm/radeon_drm.h>
+#include "radeon_reg.h"
+#include "radeon.h"
+
+
+/* Test BO GTT->VRAM and VRAM->GTT GPU copies across the whole GTT aperture */
+void radeon_test_moves(struct radeon_device *rdev)
+{
+        struct radeon_object *vram_obj = NULL;
+        struct radeon_object **gtt_obj = NULL;
+        struct radeon_fence *fence = NULL;
+        uint64_t gtt_addr, vram_addr;
+        unsigned i, n, size;
+        int r;
+
+        size = 1024 * 1024;
+
+        /* Number of tests =
+         * (Total GTT - IB pool - writeback page - ring buffer) / test size
+         */
+        n = (rdev->mc.gtt_size - RADEON_IB_POOL_SIZE*64*1024 - 4096 -
+             rdev->cp.ring_size) / size;
+
+        gtt_obj = kzalloc(n * sizeof(*gtt_obj), GFP_KERNEL);
+        if (!gtt_obj) {
+                DRM_ERROR("Failed to allocate %d pointers\n", n);
+                r = 1;
+                goto out_cleanup;
+        }
+
+        r = radeon_object_create(rdev, NULL, size, true, RADEON_GEM_DOMAIN_VRAM,
+                                 false, &vram_obj);
+        if (r) {
+                DRM_ERROR("Failed to create VRAM object\n");
+                goto out_cleanup;
+        }
+
+        r = radeon_object_pin(vram_obj, RADEON_GEM_DOMAIN_VRAM, &vram_addr);
+        if (r) {
+                DRM_ERROR("Failed to pin VRAM object\n");
+                goto out_cleanup;
+        }
+
+        for (i = 0; i < n; i++) {
+                void *gtt_map, *vram_map;
+                void **gtt_start, **gtt_end;
+                void **vram_start, **vram_end;
+
+                r = radeon_object_create(rdev, NULL, size, true,
+                                         RADEON_GEM_DOMAIN_GTT, false, gtt_obj + i);
+                if (r) {
+                        DRM_ERROR("Failed to create GTT object %d\n", i);
+                        goto out_cleanup;
+                }
+
+                r = radeon_object_pin(gtt_obj[i], RADEON_GEM_DOMAIN_GTT, &gtt_addr);
+                if (r) {
+                        DRM_ERROR("Failed to pin GTT object %d\n", i);
+                        goto out_cleanup;
+                }
+
+                r = radeon_object_kmap(gtt_obj[i], &gtt_map);
+                if (r) {
+                        DRM_ERROR("Failed to map GTT object %d\n", i);
+                        goto out_cleanup;
+                }
+
+                for (gtt_start = gtt_map, gtt_end = gtt_map + size;
+                     gtt_start < gtt_end;
+                     gtt_start++)
+                        *gtt_start = gtt_start;
+
+                radeon_object_kunmap(gtt_obj[i]);
+
+                r = radeon_fence_create(rdev, &fence);
+                if (r) {
+                        DRM_ERROR("Failed to create GTT->VRAM fence %d\n", i);
+                        goto out_cleanup;
+                }
+
+                r = radeon_copy(rdev, gtt_addr, vram_addr, size / 4096, fence);
+                if (r) {
+                        DRM_ERROR("Failed GTT->VRAM copy %d\n", i);
+                        goto out_cleanup;
+                }
+
+                r = radeon_fence_wait(fence, false);
+                if (r) {
+                        DRM_ERROR("Failed to wait for GTT->VRAM fence %d\n", i);
+                        goto out_cleanup;
+                }
+
+                radeon_fence_unref(&fence);
+
+                r = radeon_object_kmap(vram_obj, &vram_map);
+                if (r) {
+                        DRM_ERROR("Failed to map VRAM object after copy %d\n", i);
+                        goto out_cleanup;
+                }
+
+                for (gtt_start = gtt_map, gtt_end = gtt_map + size,
+                     vram_start = vram_map, vram_end = vram_map + size;
+                     vram_start < vram_end;
+                     gtt_start++, vram_start++) {
+                        if (*vram_start != gtt_start) {
+                                DRM_ERROR("Incorrect GTT->VRAM copy %d: Got 0x%p, "
+                                          "expected 0x%p (GTT map 0x%p-0x%p)\n",
+                                          i, *vram_start, gtt_start, gtt_map,
+                                          gtt_end);
+                                radeon_object_kunmap(vram_obj);
+                                goto out_cleanup;
+                        }
+                        *vram_start = vram_start;
+                }
+
+                radeon_object_kunmap(vram_obj);
+
+                r = radeon_fence_create(rdev, &fence);
+                if (r) {
+                        DRM_ERROR("Failed to create VRAM->GTT fence %d\n", i);
+                        goto out_cleanup;
+                }
+
+                r = radeon_copy(rdev, vram_addr, gtt_addr, size / 4096, fence);
+                if (r) {
+                        DRM_ERROR("Failed VRAM->GTT copy %d\n", i);
+                        goto out_cleanup;
+                }
+
+                r = radeon_fence_wait(fence, false);
+                if (r) {
+                        DRM_ERROR("Failed to wait for VRAM->GTT fence %d\n", i);
+                        goto out_cleanup;
+                }
+
+                radeon_fence_unref(&fence);
+
+                r = radeon_object_kmap(gtt_obj[i], &gtt_map);
+                if (r) {
+                        DRM_ERROR("Failed to map GTT object after copy %d\n", i);
+                        goto out_cleanup;
+                }
+
+                for (gtt_start = gtt_map, gtt_end = gtt_map + size,
+                     vram_start = vram_map, vram_end = vram_map + size;
+                     gtt_start < gtt_end;
+                     gtt_start++, vram_start++) {
+                        if (*gtt_start != vram_start) {
+                                DRM_ERROR("Incorrect VRAM->GTT copy %d: Got 0x%p, "
+                                          "expected 0x%p (VRAM map 0x%p-0x%p)\n",
+                                          i, *gtt_start, vram_start, vram_map,
+                                          vram_end);
+                                radeon_object_kunmap(gtt_obj[i]);
+                                goto out_cleanup;
+                        }
+                }
+
+                radeon_object_kunmap(gtt_obj[i]);
+
+                DRM_INFO("Tested GTT->VRAM and VRAM->GTT copy for GTT offset 0x%llx\n",
+                         gtt_addr - rdev->mc.gtt_location);
+        }
+
+out_cleanup:
+        if (vram_obj) {
+                radeon_object_unpin(vram_obj);
+                radeon_object_unref(&vram_obj);
+        }
+        if (gtt_obj) {
+                for (i = 0; i < n; i++) {
+                        if (gtt_obj[i]) {
+                                radeon_object_unpin(gtt_obj[i]);
+                                radeon_object_unref(&gtt_obj[i]);
+                        }
+                }
+                kfree(gtt_obj);
+        }
+        if (fence) {
+                radeon_fence_unref(&fence);
+        }
+        if (r) {
+                printk(KERN_WARNING "Error while testing BO move.\n");
+        }
+}
+