tests: move xglinfo to demos

2 files changed, 640 insertions, 0 deletions

diff --git a/demos/CMakeLists.txt b/demos/CMakeLists.txt
index 33a3f8ac..40488f29 100644
--- a/demos/CMakeLists.txt
+++ b/demos/CMakeLists.txt
@@ -34,6 +34,8 @@ include_directories (
 link_directories(${XCB_LIBRARY_DIRS})
 link_libraries(${XCB_LIBRARIES} XGL png m)
 
+add_executable(xglinfo xglinfo.c)
+
 add_executable(tri tri.c)
 target_link_libraries(tri)
 
diff --git a/demos/xglinfo.c b/demos/xglinfo.c
new file mode 100644
index 00000000..4bbd635c
--- /dev/null
+++ b/demos/xglinfo.c
@@ -0,0 +1,638 @@
+/*
+ * XGL
+ *
+ * Copyright (C) 2014 LunarG, 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 AUTHORS OR COPYRIGHT HOLDERS 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.
+ */
+#include <stdlib.h>
+#include <stdio.h>
+#include <stdbool.h>
+#include <string.h>
+#include <assert.h>
+
+#include <xgl.h>
+
+#define ERR(err) printf("%s:%d: failed with %s\n", \
+    __FILE__, __LINE__, xgl_result_string(err));
+
+#define ERR_EXIT(err) do { ERR(err); exit(-1); } while (0)
+
+#define ARRAY_SIZE(a) (sizeof(a) / sizeof(a[0]))
+
+#define MAX_GPUS 8
+
+#define MAX_QUEUE_TYPES 5
+
+struct app_gpu;
+
+struct app_dev {
+    struct app_gpu *gpu; /* point back to the GPU */
+
+    XGL_DEVICE obj;
+
+    XGL_UINT heap_count;
+    XGL_MEMORY_HEAP_PROPERTIES *heap_props;
+
+    XGL_FORMAT_PROPERTIES format_props[XGL_MAX_CH_FMT][XGL_MAX_NUM_FMT];
+};
+
+struct app_gpu {
+    XGL_UINT id;
+    XGL_PHYSICAL_GPU obj;
+
+    XGL_PHYSICAL_GPU_PROPERTIES props;
+    XGL_PHYSICAL_GPU_PERFORMANCE perf;
+
+    XGL_UINT queue_count;
+    XGL_PHYSICAL_GPU_QUEUE_PROPERTIES *queue_props;
+    XGL_DEVICE_QUEUE_CREATE_INFO *queue_reqs;
+
+    XGL_PHYSICAL_GPU_MEMORY_PROPERTIES memory_props;
+
+    XGL_UINT extension_count;
+    const XGL_CHAR **extensions;
+
+    struct app_dev dev;
+};
+
+static const char *xgl_result_string(XGL_RESULT err)
+{
+    switch (err) {
+#define STR(r) case r: return #r
+    STR(XGL_SUCCESS);
+    STR(XGL_UNSUPPORTED);
+    STR(XGL_NOT_READY);
+    STR(XGL_TIMEOUT);
+    STR(XGL_EVENT_SET);
+    STR(XGL_EVENT_RESET);
+    STR(XGL_ERROR_UNKNOWN);
+    STR(XGL_ERROR_UNAVAILABLE);
+    STR(XGL_ERROR_INITIALIZATION_FAILED);
+    STR(XGL_ERROR_OUT_OF_MEMORY);
+    STR(XGL_ERROR_OUT_OF_GPU_MEMORY);
+    STR(XGL_ERROR_DEVICE_ALREADY_CREATED);
+    STR(XGL_ERROR_DEVICE_LOST);
+    STR(XGL_ERROR_INVALID_POINTER);
+    STR(XGL_ERROR_INVALID_VALUE);
+    STR(XGL_ERROR_INVALID_HANDLE);
+    STR(XGL_ERROR_INVALID_ORDINAL);
+    STR(XGL_ERROR_INVALID_MEMORY_SIZE);
+    STR(XGL_ERROR_INVALID_EXTENSION);
+    STR(XGL_ERROR_INVALID_FLAGS);
+    STR(XGL_ERROR_INVALID_ALIGNMENT);
+    STR(XGL_ERROR_INVALID_FORMAT);
+    STR(XGL_ERROR_INVALID_IMAGE);
+    STR(XGL_ERROR_INVALID_DESCRIPTOR_SET_DATA);
+    STR(XGL_ERROR_INVALID_QUEUE_TYPE);
+    STR(XGL_ERROR_INVALID_OBJECT_TYPE);
+    STR(XGL_ERROR_UNSUPPORTED_SHADER_IL_VERSION);
+    STR(XGL_ERROR_BAD_SHADER_CODE);
+    STR(XGL_ERROR_BAD_PIPELINE_DATA);
+    STR(XGL_ERROR_TOO_MANY_MEMORY_REFERENCES);
+    STR(XGL_ERROR_NOT_MAPPABLE);
+    STR(XGL_ERROR_MEMORY_MAP_FAILED);
+    STR(XGL_ERROR_MEMORY_UNMAP_FAILED);
+    STR(XGL_ERROR_INCOMPATIBLE_DEVICE);
+    STR(XGL_ERROR_INCOMPATIBLE_DRIVER);
+    STR(XGL_ERROR_INCOMPLETE_COMMAND_BUFFER);
+    STR(XGL_ERROR_BUILDING_COMMAND_BUFFER);
+    STR(XGL_ERROR_MEMORY_NOT_BOUND);
+    STR(XGL_ERROR_INCOMPATIBLE_QUEUE);
+    STR(XGL_ERROR_NOT_SHAREABLE);
+#undef STR
+    default: return "UNKNOWN_RESULT";
+    }
+}
+
+static const char *xgl_gpu_type_string(XGL_PHYSICAL_GPU_TYPE type)
+{
+    switch (type) {
+#define STR(r) case XGL_GPU_TYPE_ ##r: return #r
+    STR(OTHER);
+    STR(INTEGRATED);
+    STR(DISCRETE);
+    STR(VIRTUAL);
+#undef STR
+    default: return "UNKNOWN_GPU";
+    }
+}
+
+static const char *xgl_heap_type_string(XGL_HEAP_MEMORY_TYPE type)
+{
+    switch (type) {
+#define STR(r) case XGL_HEAP_MEMORY_ ##r: return #r
+    STR(OTHER);
+    STR(LOCAL);
+    STR(REMOTE);
+    STR(EMBEDDED);
+#undef STR
+    default: return "UNKNOWN_HEAP";
+    }
+}
+
+static const char *xgl_channel_format_string(XGL_CHANNEL_FORMAT ch)
+{
+    switch (ch) {
+#define STR(r) case XGL_CH_FMT_ ##r: return #r
+    STR(UNDEFINED);
+    STR(R4G4);
+    STR(R4G4B4A4);
+    STR(R5G6B5);
+    STR(B5G6R5);
+    STR(R5G5B5A1);
+    STR(R8);
+    STR(R8G8);
+    STR(R8G8B8A8);
+    STR(B8G8R8A8);
+    STR(R10G11B11);
+    STR(R11G11B10);
+    STR(R10G10B10A2);
+    STR(R16);
+    STR(R16G16);
+    STR(R16G16B16A16);
+    STR(R32);
+    STR(R32G32);
+    STR(R32G32B32);
+    STR(R32G32B32A32);
+    STR(R16G8);
+    STR(R32G8);
+    STR(R9G9B9E5);
+    STR(BC1);
+    STR(BC2);
+    STR(BC3);
+    STR(BC4);
+    STR(BC5);
+    STR(BC6U);
+    STR(BC6S);
+    STR(BC7);
+#undef STR
+    default: return "UNKNOWN_CH";
+    }
+}
+
+static const char *xgl_numeric_format_string(XGL_NUM_FORMAT num)
+{
+    switch (num) {
+#define STR(r) case XGL_NUM_FMT_ ##r: return #r
+    STR(UNDEFINED);
+    STR(UNORM);
+    STR(SNORM);
+    STR(UINT);
+    STR(SINT);
+    STR(FLOAT);
+    STR(SRGB);
+    STR(DS);
+#undef STR
+    default: return "UNKNOWN_NUM";
+    }
+}
+
+static void app_dev_init_formats(struct app_dev *dev)
+{
+    XGL_CHANNEL_FORMAT ch;
+    XGL_NUM_FORMAT num;
+
+    for (ch = 0; ch < XGL_MAX_CH_FMT; ch++) {
+        for (num = 0; num < XGL_MAX_NUM_FMT; num++) {
+            const XGL_FORMAT fmt = {
+                .channelFormat = ch,
+                .numericFormat = num,
+            };
+            XGL_RESULT err;
+            XGL_SIZE size = sizeof(dev->format_props[ch][num]);
+
+            err = xglGetFormatInfo(dev->obj, fmt,
+                                   XGL_INFO_TYPE_FORMAT_PROPERTIES,
+                                   &size, &dev->format_props[ch][num]);
+            if (err) {
+                memset(&dev->format_props[ch][num], 0,
+                       sizeof(dev->format_props[ch][num]));
+            }
+            else if (size != sizeof(dev->format_props[ch][num])) {
+                ERR_EXIT(XGL_ERROR_UNKNOWN);
+            }
+        }
+    }
+}
+
+static void app_dev_init(struct app_dev *dev, struct app_gpu *gpu)
+{
+    XGL_DEVICE_CREATE_INFO info = {
+        .sType = XGL_STRUCTURE_TYPE_DEVICE_CREATE_INFO,
+        .pNext = NULL,
+        .queueRecordCount = 0,
+        .pRequestedQueues = NULL,
+        .extensionCount = 0,
+        .ppEnabledExtensionNames = NULL,
+        .maxValidationLevel = XGL_VALIDATION_LEVEL_END_RANGE,
+        .flags = XGL_DEVICE_CREATE_VALIDATION_BIT,
+    };
+    XGL_RESULT err;
+    XGL_SIZE size;
+    XGL_UINT i;
+
+    /* request all queues */
+    info.queueRecordCount = gpu->queue_count;
+    info.pRequestedQueues = gpu->queue_reqs;
+
+    /* enable all extensions */
+    info.extensionCount = gpu->extension_count;
+    info.ppEnabledExtensionNames = gpu->extensions;
+    dev->gpu = gpu;
+    err = xglCreateDevice(gpu->obj, &info, &dev->obj);
+    if (err)
+        ERR_EXIT(err);
+
+    err = xglGetMemoryHeapCount(dev->obj, &dev->heap_count);
+    if (err)
+        ERR_EXIT(err);
+
+    dev->heap_props =
+            malloc(sizeof(dev->heap_props[0]) * dev->heap_count);
+    if (!dev->heap_props)
+        ERR_EXIT(XGL_ERROR_OUT_OF_MEMORY);
+
+    for (i = 0; i < dev->heap_count; i++) {
+        size = sizeof(dev->heap_props[0]);
+        err = xglGetMemoryHeapInfo(dev->obj, i,
+                                   XGL_INFO_TYPE_MEMORY_HEAP_PROPERTIES,
+                                   &size, &dev->heap_props[i]);
+        if (err || size != sizeof(dev->heap_props[0]))
+            ERR_EXIT(err);
+    }
+}
+
+static void app_dev_destroy(struct app_dev *dev)
+{
+    free(dev->heap_props);
+    xglDestroyDevice(dev->obj);
+}
+
+static void app_gpu_init_extensions(struct app_gpu *gpu)
+{
+    XGL_RESULT err;
+    XGL_UINT i;
+
+    static const XGL_CHAR *known_extensions[] = {
+        (const XGL_CHAR *) "XGL_WSI_X11",
+    };
+
+    for (i = 0; i < ARRAY_SIZE(known_extensions); i++) {
+        err = xglGetExtensionSupport(gpu->obj, known_extensions[i]);
+        if (!err)
+            gpu->extension_count++;
+    }
+
+    gpu->extensions =
+            malloc(sizeof(gpu->extensions[0]) * gpu->extension_count);
+    if (!gpu->extensions)
+        ERR_EXIT(XGL_ERROR_OUT_OF_MEMORY);
+
+    gpu->extension_count = 0;
+    for (i = 0; i < ARRAY_SIZE(known_extensions); i++) {
+        err = xglGetExtensionSupport(gpu->obj, known_extensions[i]);
+        if (!err)
+            gpu->extensions[gpu->extension_count++] = known_extensions[i];
+    }
+}
+
+static void app_gpu_init(struct app_gpu *gpu, XGL_UINT id, XGL_PHYSICAL_GPU obj)
+{
+    XGL_SIZE size;
+    XGL_RESULT err;
+    int i;
+
+    memset(gpu, 0, sizeof(*gpu));
+
+    gpu->id = id;
+    gpu->obj = obj;
+    size = sizeof(gpu->props);
+    err = xglGetGpuInfo(gpu->obj,
+                        XGL_INFO_TYPE_PHYSICAL_GPU_PROPERTIES,
+                        &size, &gpu->props);
+    if (err || size != sizeof(gpu->props))
+        ERR_EXIT(err);
+
+    size = sizeof(gpu->perf);
+    err = xglGetGpuInfo(gpu->obj,
+                        XGL_INFO_TYPE_PHYSICAL_GPU_PERFORMANCE,
+                        &size, &gpu->perf);
+    if (err || size != sizeof(gpu->perf))
+        ERR_EXIT(err);
+
+    /* get queue count */
+    err = xglGetGpuInfo(gpu->obj,
+                        XGL_INFO_TYPE_PHYSICAL_GPU_QUEUE_PROPERTIES,
+                        &size, NULL);
+    if (err || size % sizeof(gpu->queue_props[0]))
+        ERR_EXIT(err);
+    gpu->queue_count = size / sizeof(gpu->queue_props[0]);
+
+    gpu->queue_props =
+            malloc(sizeof(gpu->queue_props[0]) * gpu->queue_count);
+    size = sizeof(gpu->queue_props[0]) * gpu->queue_count;
+    if (!gpu->queue_props)
+        ERR_EXIT(XGL_ERROR_OUT_OF_MEMORY);
+    err = xglGetGpuInfo(gpu->obj,
+                        XGL_INFO_TYPE_PHYSICAL_GPU_QUEUE_PROPERTIES,
+                        &size, gpu->queue_props);
+    if (err || size != sizeof(gpu->queue_props[0]) * gpu->queue_count)
+        ERR_EXIT(err);
+
+    /* set up queue requests */
+    size = sizeof(*gpu->queue_reqs) * gpu->queue_count;
+    gpu->queue_reqs = malloc(sizeof(*gpu->queue_reqs) * gpu->queue_count);
+    if (!gpu->queue_reqs)
+        ERR_EXIT(XGL_ERROR_OUT_OF_MEMORY);
+    for (i = 0; i < gpu->queue_count; i++) {
+        gpu->queue_reqs[i].queueNodeIndex = i;
+        gpu->queue_reqs[i].queueCount = gpu->queue_props[i].queueCount;
+    }
+
+    size = sizeof(gpu->memory_props);
+    err = xglGetGpuInfo(gpu->obj,
+                        XGL_INFO_TYPE_PHYSICAL_GPU_MEMORY_PROPERTIES,
+                        &size, &gpu->memory_props);
+    if (err || size != sizeof(gpu->memory_props))
+        ERR_EXIT(err);
+
+    app_gpu_init_extensions(gpu);
+    app_dev_init(&gpu->dev, gpu);
+    app_dev_init_formats(&gpu->dev);
+}
+
+static void app_gpu_destroy(struct app_gpu *gpu)
+{
+    app_dev_destroy(&gpu->dev);
+    free(gpu->extensions);
+    free(gpu->queue_reqs);
+    free(gpu->queue_props);
+}
+
+static void app_dev_dump_format_props(const struct app_dev *dev, XGL_CHANNEL_FORMAT ch, XGL_NUM_FORMAT num)
+{
+    const XGL_FORMAT_PROPERTIES *props = &dev->format_props[ch][num];
+    struct {
+        const char *name;
+        XGL_FLAGS flags;
+    } tilings[2];
+    XGL_UINT i;
+
+    if (!props->linearTilingFeatures && !props->optimalTilingFeatures)
+        return;
+
+    tilings[0].name = "linear";
+    tilings[0].flags = props->linearTilingFeatures;
+    tilings[1].name = "optimal";
+    tilings[1].flags = props->optimalTilingFeatures;
+
+    printf("FORMAT_%s_%s\n", xgl_channel_format_string(ch),
+            xgl_numeric_format_string(num));
+    for (i = 0; i < ARRAY_SIZE(tilings); i++) {
+        if (!tilings[i].flags)
+            continue;
+
+        printf("\t%s tiling image =%s%s%s\n", tilings[i].name,
+                (tilings[i].flags & XGL_FORMAT_IMAGE_SHADER_READ_BIT)      ? " read" : "",
+                (tilings[i].flags & XGL_FORMAT_IMAGE_SHADER_WRITE_BIT)     ? " write" : "",
+                (tilings[i].flags & XGL_FORMAT_IMAGE_COPY_BIT)             ? " copy" : "");
+        printf("\t%s tiling memory =%s\n", tilings[i].name,
+                (tilings[i].flags & XGL_FORMAT_MEMORY_SHADER_ACCESS_BIT)   ? " access" : "");
+        printf("\t%s tiling attachment =%s%s%s%s%s\n", tilings[i].name,
+                (tilings[i].flags & XGL_FORMAT_COLOR_ATTACHMENT_WRITE_BIT) ? " color" : "",
+                (tilings[i].flags & XGL_FORMAT_COLOR_ATTACHMENT_BLEND_BIT) ? " blend" : "",
+                (tilings[i].flags & XGL_FORMAT_DEPTH_ATTACHMENT_BIT)       ? " depth" : "",
+                (tilings[i].flags & XGL_FORMAT_STENCIL_ATTACHMENT_BIT)     ? " stencil" : "",
+                (tilings[i].flags & XGL_FORMAT_MSAA_ATTACHMENT_BIT)        ? " msaa" : "");
+        printf("\t%s tiling conversion = %u\n", tilings[i].name,
+                (bool) (tilings[i].flags & XGL_FORMAT_CONVERSION_BIT));
+    }
+}
+
+static void app_dev_dump_heap_props(const struct app_dev *dev, XGL_UINT id)
+{
+    const XGL_MEMORY_HEAP_PROPERTIES *props = &dev->heap_props[id];
+
+    printf("XGL_MEMORY_HEAP_PROPERTIES[%u]\n", id);
+    printf("\tstructSize = %u\n",       props->structSize);
+    printf("\theapMemoryType = %s\n",   xgl_heap_type_string(props->heapMemoryType));
+    printf("\theapSize = %u\n",         props->heapSize);
+    printf("\tpagesSize = %u\n",        props->pageSize);
+
+    printf("\tflags =%s%s%s%s%s%s\n",
+            (props->flags & XGL_MEMORY_HEAP_CPU_VISIBLE_BIT)        ? " visible" : "",
+            (props->flags & XGL_MEMORY_HEAP_CPU_GPU_COHERENT_BIT)   ? " coherent" : "",
+            (props->flags & XGL_MEMORY_HEAP_CPU_UNCACHED_BIT)       ? " uc" : "",
+            (props->flags & XGL_MEMORY_HEAP_CPU_WRITE_COMBINED_BIT) ? " wc" : "",
+            (props->flags & XGL_MEMORY_HEAP_HOLDS_PINNED_BIT)       ? " pinnable" : "",
+            (props->flags & XGL_MEMORY_HEAP_SHAREABLE_BIT)          ? " shareable" : "");
+
+    printf("\tgpuReadPerfRating = %f\n",    props->gpuReadPerfRating);
+    printf("\tgpuWritePerfRating = %f\n",   props->gpuWritePerfRating);
+    printf("\tcpuReadPerfRating = %f\n",    props->cpuReadPerfRating);
+    printf("\tcpuWritePerfRating = %f\n",   props->cpuWritePerfRating);
+}
+
+static void
+app_dev_dump(const struct app_dev *dev)
+{
+    XGL_CHANNEL_FORMAT ch;
+    XGL_NUM_FORMAT num;
+    XGL_UINT i;
+
+    for (i = 0; i < dev->heap_count; i++) {
+        app_dev_dump_heap_props(dev, i);
+        printf("\n");
+    }
+
+    for (ch = 0; ch < XGL_MAX_CH_FMT; ch++) {
+        for (num = 0; num < XGL_MAX_NUM_FMT; num++)
+            app_dev_dump_format_props(dev, ch, num);
+    }
+}
+
+static void app_gpu_dump_multi_compat(const struct app_gpu *gpu, const struct app_gpu *other,
+        const XGL_GPU_COMPATIBILITY_INFO *info)
+{
+    printf("XGL_GPU_COMPATIBILITY_INFO[GPU%d]\n", other->id);
+
+#define TEST(info, b) printf(#b " = %u\n", (bool) (info->compatibilityFlags & XGL_GPU_COMPAT_ ##b## _BIT))
+    TEST(info, ASIC_FEATURES);
+    TEST(info, IQ_MATCH);
+    TEST(info, PEER_TRANSFER);
+    TEST(info, SHARED_MEMORY);
+    TEST(info, SHARED_SYNC);
+    TEST(info, SHARED_GPU0_DISPLAY);
+    TEST(info, SHARED_GPU1_DISPLAY);
+#undef TEST
+}
+
+static void app_gpu_multi_compat(struct app_gpu *gpus, XGL_UINT gpu_count)
+{
+        XGL_RESULT err;
+        XGL_UINT i, j;
+
+        for (i = 0; i < gpu_count; i++) {
+                for (j = 0; j < gpu_count; j++) {
+                        XGL_GPU_COMPATIBILITY_INFO info;
+
+                        if (i == j)
+                                continue;
+
+                        err = xglGetMultiGpuCompatibility(gpus[i].obj,
+                                        gpus[j].obj, &info);
+                        if (err)
+                                ERR_EXIT(err);
+
+                        app_gpu_dump_multi_compat(&gpus[i], &gpus[j], &info);
+                }
+        }
+}
+
+static void app_gpu_dump_props(const struct app_gpu *gpu)
+{
+    const XGL_PHYSICAL_GPU_PROPERTIES *props = &gpu->props;
+
+    printf("XGL_PHYSICAL_GPU_PROPERTIES\n");
+    printf("\tstructSize = %u\n",                   props->structSize);
+    printf("\tapiVersion = %u\n",                   props->apiVersion);
+    printf("\tdriverVersion = %u\n",                props->driverVersion);
+    printf("\tvendorId = 0x%04x\n",                 props->vendorId);
+    printf("\tdeviceId = 0x%04x\n",                 props->deviceId);
+    printf("\tgpuType = %s\n",                      xgl_gpu_type_string(props->gpuType));
+    printf("\tgpuName = %s\n",                      props->gpuName);
+    printf("\tmaxMemRefsPerSubmission = %u\n",      props->maxMemRefsPerSubmission);
+    printf("\tvirtualMemPageSize = %u\n",           props->virtualMemPageSize);
+    printf("\tmaxInlineMemoryUpdateSize = %u\n",    props->maxInlineMemoryUpdateSize);
+    printf("\tmaxBoundDescriptorSets = %u\n",       props->maxBoundDescriptorSets);
+    printf("\tmaxThreadGroupSize = %u\n",           props->maxThreadGroupSize);
+    printf("\ttimestampFrequency = %lu\n",          props->timestampFrequency);
+    printf("\tmultiColorAttachmentClears = %u\n",   props->multiColorAttachmentClears);
+}
+
+static void app_gpu_dump_perf(const struct app_gpu *gpu)
+{
+    const XGL_PHYSICAL_GPU_PERFORMANCE *perf = &gpu->perf;
+
+    printf("XGL_PHYSICAL_GPU_PERFORMANCE\n");
+    printf("\tmaxGpuClock = %f\n",      perf->maxGpuClock);
+    printf("\taluPerClock = %f\n",      perf->aluPerClock);
+    printf("\ttexPerClock = %f\n",      perf->texPerClock);
+    printf("\tprimsPerClock = %f\n",    perf->primsPerClock);
+    printf("\tpixelsPerClock = %f\n",   perf->pixelsPerClock);
+}
+
+static void app_gpu_dump_extensions(const struct app_gpu *gpu)
+{
+    int i;
+    printf("Extensions");
+    printf("\tcount = %d\n",            gpu->extension_count);
+    printf("\t");
+    for (i=0; i< gpu->extension_count; i++) {
+        if (i>0)
+            printf(", "); // separator between extension names
+        printf("%s",                    gpu->extensions[i]);
+    }
+    printf("\n");
+}
+
+static void app_gpu_dump_queue_props(const struct app_gpu *gpu, XGL_UINT id)
+{
+    const XGL_PHYSICAL_GPU_QUEUE_PROPERTIES *props = &gpu->queue_props[id];
+
+    printf("XGL_PHYSICAL_GPU_QUEUE_PROPERTIES[%d]\n", id);
+    printf("\tstructSize = %u\n",           props->structSize);
+    printf("\tqueueFlags = %c%c%c%c\n",
+            (props->queueFlags & XGL_QUEUE_GRAPHICS_BIT) ? 'G' : '.',
+            (props->queueFlags & XGL_QUEUE_COMPUTE_BIT)  ? 'C' : '.',
+            (props->queueFlags & XGL_QUEUE_DMA_BIT)      ? 'D' : '.',
+            (props->queueFlags & XGL_QUEUE_EXTENDED_BIT) ? 'X' : '.');
+    printf("\tqueueCount = %u\n",           props->queueCount);
+    printf("\tmaxAtomicCounters = %u\n",    props->maxAtomicCounters);
+    printf("\tsupportsTimestamps = %u\n",   props->supportsTimestamps);
+}
+
+static void app_gpu_dump_memory_props(const struct app_gpu *gpu)
+{
+    const XGL_PHYSICAL_GPU_MEMORY_PROPERTIES *props = &gpu->memory_props;
+
+    printf("XGL_PHYSICAL_GPU_MEMORY_PROPERTIES\n");
+    printf("\tstructSize = %u\n",                       props->structSize);
+    printf("\tsupportsMigration = %u\n",                props->supportsMigration);
+    printf("\tsupportsVirtualMemoryRemapping = %u\n",   props->supportsVirtualMemoryRemapping);
+    printf("\tsupportsPinning = %u\n",                  props->supportsPinning);
+}
+
+static void app_gpu_dump(const struct app_gpu *gpu)
+{
+    XGL_UINT i;
+
+    printf("GPU%u\n", gpu->id);
+    app_gpu_dump_props(gpu);
+    printf("\n");
+    app_gpu_dump_extensions(gpu);
+    printf("\n");
+    app_gpu_dump_perf(gpu);
+    printf("\n");
+    for (i = 0; i < gpu->queue_count; i++) {
+        app_gpu_dump_queue_props(gpu, i);
+        printf("\n");
+    }
+    app_gpu_dump_memory_props(gpu);
+    printf("\n");
+    app_dev_dump(&gpu->dev);
+}
+
+int main(int argc, char **argv)
+{
+    static const XGL_APPLICATION_INFO app_info = {
+        .sType = XGL_STRUCTURE_TYPE_APPLICATION_INFO,
+        .pNext = NULL,
+        .pAppName = (const XGL_CHAR *) "xglinfo",
+        .appVersion = 1,
+        .pEngineName = (const XGL_CHAR *) "xglinfo",
+        .engineVersion = 1,
+        .apiVersion = XGL_MAKE_VERSION(0, 22, 0),
+    };
+    struct app_gpu gpus[MAX_GPUS];
+    XGL_PHYSICAL_GPU objs[MAX_GPUS];
+    XGL_UINT gpu_count, i;
+    XGL_RESULT err;
+
+    err = xglInitAndEnumerateGpus(&app_info, NULL,
+            MAX_GPUS, &gpu_count, objs);
+    if (err)
+        ERR_EXIT(err);
+
+    for (i = 0; i < gpu_count; i++) {
+        app_gpu_init(&gpus[i], i, objs[i]);
+        app_gpu_dump(&gpus[i]);
+        printf("\n\n");
+    }
+
+    app_gpu_multi_compat(gpus, gpu_count);
+
+    for (i = 0; i < gpu_count; i++)
+        app_gpu_destroy(&gpus[i]);
+
+    xglInitAndEnumerateGpus(&app_info, NULL, 0, &gpu_count, NULL);
+
+    return 0;
+}