GPU (Linux): refactor; add amdgpu DRM connector

This commit is contained in:
Carter Li
2024-09-23 14:32:38 +08:00
parent e0296b75f6
commit b3b6e5fec1
3 changed files with 248 additions and 120 deletions
+2 -2
View File
@@ -7,7 +7,7 @@
#define FF_GPU_VMEM_SIZE_UNSET ((uint64_t)-1)
#define FF_GPU_FREQUENCY_UNSET 0
#define FF_GPU_CORE_USAGE_UNSET (0/0.0)
#define FF_GPU_INDEX_UNSET ((uint8_t)-1)
#define FF_GPU_INDEX_UNSET ((uint32_t)-1)
extern const char* FF_GPU_VENDOR_NAME_APPLE;
extern const char* FF_GPU_VENDOR_NAME_AMD;
@@ -40,7 +40,7 @@ typedef struct FFGPUResult
uint32_t frequency; // Maximum time clock frequency in MHz
FFGPUMemory dedicated;
FFGPUMemory shared;
uint64_t deviceId; // Used internally, may be uninitialized
uint64_t deviceId;
} FFGPUResult;
const char* ffDetectGPU(const FFGPUOptions* options, FFlist* result);
+91
View File
@@ -0,0 +1,91 @@
#if __aarch64__ && FF_HAVE_DRM
#include <drm.h>
#include <fcntl.h>
#include <sys/ioctl.h>
#if __has_include(<drm/asahi_drm.h>)
#include <drm/asahi_drm.h>
#else
// https://github.com/AsahiLinux/linux/blob/asahi/include/uapi/drm/asahi_drm.h
/* SPDX-License-Identifier: MIT */
/* Copyright (C) The Asahi Linux Contributors */
#define DRM_ASAHI_UNSTABLE_UABI_VERSION 10011
#define DRM_ASAHI_GET_PARAMS 0x00
#define DRM_ASAHI_MAX_CLUSTERS 32
struct drm_asahi_params_global
{
__u32 unstable_uabi_version;
__u32 pad0;
__u64 feat_compat;
__u64 feat_incompat;
__u32 gpu_generation;
__u32 gpu_variant;
__u32 gpu_revision;
__u32 chip_id;
__u32 num_dies;
__u32 num_clusters_total;
__u32 num_cores_per_cluster;
__u32 num_frags_per_cluster;
__u32 num_gps_per_cluster;
__u32 num_cores_total_active;
__u64 core_masks[DRM_ASAHI_MAX_CLUSTERS];
__u32 vm_page_size;
__u32 pad1;
__u64 vm_user_start;
__u64 vm_user_end;
__u64 vm_usc_start;
__u64 vm_usc_end;
__u64 vm_kernel_min_size;
__u32 max_syncs_per_submission;
__u32 max_commands_per_submission;
__u32 max_commands_in_flight;
__u32 max_attachments;
__u32 timer_frequency_hz;
__u32 min_frequency_khz;
__u32 max_frequency_khz;
__u32 max_power_mw;
__u32 result_render_size;
__u32 result_compute_size;
__u32 firmware_version[4];
};
struct drm_asahi_get_params
{
/** @extensions: Pointer to the first extension struct, if any */
__u64 extensions;
/** @param: Parameter group to fetch (MBZ) */
__u32 param_group;
/** @pad: MBZ */
__u32 pad;
/** @value: User pointer to write parameter struct */
__u64 pointer;
/** @value: Size of user buffer, max size supported on return */
__u64 size;
};
enum
{
DRM_IOCTL_ASAHI_GET_PARAMS = DRM_IOWR(DRM_COMMAND_BASE + DRM_ASAHI_GET_PARAMS, struct drm_asahi_get_params),
};
#endif // __has_include
#endif // FF_HAVE_DRM
+155 -118
View File
@@ -4,97 +4,24 @@
#include "detection/cpu/cpu.h"
#include "detection/gpu/gpu_driver_specific.h"
#include "common/io/io.h"
#include "common/library.h"
#include "common/properties.h"
#include "util/stringUtils.h"
#include <inttypes.h>
#if __aarch64__ && FF_HAVE_DRM
#include <drm.h>
#ifdef FF_HAVE_DRM
#include <amdgpu.h>
#include <amdgpu_drm.h>
#include <fcntl.h>
#include <sys/ioctl.h>
// https://github.com/AsahiLinux/linux/blob/asahi/include/uapi/drm/asahi_drm.h
/* SPDX-License-Identifier: MIT */
/* Copyright (C) The Asahi Linux Contributors */
#define DRM_ASAHI_UNSTABLE_UABI_VERSION 10011
#define DRM_ASAHI_GET_PARAMS 0x00
#define DRM_ASAHI_MAX_CLUSTERS 32
struct drm_asahi_params_global
{
__u32 unstable_uabi_version;
__u32 pad0;
__u64 feat_compat;
__u64 feat_incompat;
__u32 gpu_generation;
__u32 gpu_variant;
__u32 gpu_revision;
__u32 chip_id;
__u32 num_dies;
__u32 num_clusters_total;
__u32 num_cores_per_cluster;
__u32 num_frags_per_cluster;
__u32 num_gps_per_cluster;
__u32 num_cores_total_active;
__u64 core_masks[DRM_ASAHI_MAX_CLUSTERS];
__u32 vm_page_size;
__u32 pad1;
__u64 vm_user_start;
__u64 vm_user_end;
__u64 vm_usc_start;
__u64 vm_usc_end;
__u64 vm_kernel_min_size;
__u32 max_syncs_per_submission;
__u32 max_commands_per_submission;
__u32 max_commands_in_flight;
__u32 max_attachments;
__u32 timer_frequency_hz;
__u32 min_frequency_khz;
__u32 max_frequency_khz;
__u32 max_power_mw;
__u32 result_render_size;
__u32 result_compute_size;
__u32 firmware_version[4];
};
struct drm_asahi_get_params
{
/** @extensions: Pointer to the first extension struct, if any */
__u64 extensions;
/** @param: Parameter group to fetch (MBZ) */
__u32 param_group;
/** @pad: MBZ */
__u32 pad;
/** @value: User pointer to write parameter struct */
__u64 pointer;
/** @value: Size of user buffer, max size supported on return */
__u64 size;
};
enum
{
DRM_IOCTL_ASAHI_GET_PARAMS = DRM_IOWR(DRM_COMMAND_BASE + DRM_ASAHI_GET_PARAMS, struct drm_asahi_get_params),
};
#endif
#include "gpu_asahi.h"
#define FF_STR_INDIR(x) #x
#define FF_STR(x) FF_STR_INDIR(x)
static bool pciDetectDriver(FFGPUResult* gpu, FFstrbuf* pciDir, FFstrbuf* buffer, FF_MAYBE_UNUSED const char* drmKey)
static bool pciDetectDriver(FFstrbuf* result, FFstrbuf* pciDir, FFstrbuf* buffer, FF_MAYBE_UNUSED const char* drmKey)
{
uint32_t pciDirLength = pciDir->length;
ffStrbufAppendS(pciDir, "/driver");
@@ -106,17 +33,17 @@ static bool pciDetectDriver(FFGPUResult* gpu, FFstrbuf* pciDir, FFstrbuf* buffer
if (slash)
{
slash++;
ffStrbufSetNS(&gpu->driver, (uint32_t) (resultLength - (slash - pathBuf)), slash);
ffStrbufSetNS(result, (uint32_t) (resultLength - (slash - pathBuf)), slash);
}
if (ffStrbufEqualS(&gpu->driver, "nvidia"))
if (ffStrbufEqualS(result, "nvidia"))
{
if (ffReadFileBuffer("/proc/driver/nvidia/version", buffer))
{
if (ffStrbufContainS(buffer, " Open "))
ffStrbufAppendS(&gpu->driver, " (open source)");
ffStrbufAppendS(result, " (open source)");
else
ffStrbufAppendS(&gpu->driver, " (proprietary)");
ffStrbufAppendS(result, " (proprietary)");
}
}
@@ -126,18 +53,18 @@ static bool pciDetectDriver(FFGPUResult* gpu, FFstrbuf* pciDir, FFstrbuf* buffer
if (ffReadFileBuffer(pciDir->chars, buffer))
{
ffStrbufTrimRightSpace(buffer);
ffStrbufAppendC(&gpu->driver, ' ');
ffStrbufAppend(&gpu->driver, buffer);
ffStrbufAppendC(result, ' ');
ffStrbufAppend(result, buffer);
}
else if (ffStrbufEqualS(&gpu->driver, "zx"))
else if (ffStrbufEqualS(result, "zx"))
{
ffStrbufSubstrBefore(pciDir, pciDirLength);
ffStrbufAppendS(pciDir, "/zx_info/driver_version");
if (ffReadFileBuffer(pciDir->chars, buffer))
{
ffStrbufTrimRightSpace(buffer);
ffStrbufAppendC(&gpu->driver, ' ');
ffStrbufAppend(&gpu->driver, buffer);
ffStrbufAppendC(result, ' ');
ffStrbufAppend(result, buffer);
}
}
}
@@ -145,6 +72,78 @@ static bool pciDetectDriver(FFGPUResult* gpu, FFstrbuf* pciDir, FFstrbuf* buffer
return true;
}
static const char* drmDetectAmdSpecific(const FFGPUOptions* options, FFGPUResult* gpu, const char* drmKey, FFstrbuf* buffer)
{
#if FF_HAVE_DRM
FF_LIBRARY_LOAD(libdrm, "dlopen libdrm_amdgpu" FF_LIBRARY_EXTENSION " failed", "libdrm_amdgpu" FF_LIBRARY_EXTENSION, 2)
FF_LIBRARY_LOAD_SYMBOL_MESSAGE(libdrm, amdgpu_device_initialize)
FF_LIBRARY_LOAD_SYMBOL_MESSAGE(libdrm, amdgpu_get_marketing_name)
FF_LIBRARY_LOAD_SYMBOL_MESSAGE(libdrm, amdgpu_query_gpu_info)
FF_LIBRARY_LOAD_SYMBOL_MESSAGE(libdrm, amdgpu_query_sensor_info)
FF_LIBRARY_LOAD_SYMBOL_MESSAGE(libdrm, amdgpu_query_heap_info)
FF_LIBRARY_LOAD_SYMBOL_MESSAGE(libdrm, amdgpu_device_deinitialize)
ffStrbufSetS(buffer, "/dev/dri/");
ffStrbufAppendS(buffer, drmKey);
FF_AUTO_CLOSE_FD int fd = open(buffer->chars, O_RDONLY);
if (fd < 0) return "Failed to open DRM device";
amdgpu_device_handle handle;
uint32_t majorVersion, minorVersion;
if (ffamdgpu_device_initialize(fd, &majorVersion, &minorVersion, &handle) < 0)
return "Failed to initialize AMDGPU device";
ffStrbufAppendF(&gpu->driver, " %u.%u", (unsigned) majorVersion, (unsigned) minorVersion);
uint32_t value;
if (ffamdgpu_query_sensor_info(handle, AMDGPU_INFO_SENSOR_VDDNB, sizeof(value), &value) >= 0 && value > 0)
gpu->type = FF_GPU_TYPE_DISCRETE;
else
gpu->type = FF_GPU_TYPE_INTEGRATED;
struct amdgpu_heap_info heapInfo;
if (ffamdgpu_query_heap_info(handle, AMDGPU_GEM_DOMAIN_VRAM, 0, &heapInfo) >= 0)
{
if (gpu->type == FF_GPU_TYPE_DISCRETE)
{
gpu->dedicated.total = heapInfo.heap_size;
gpu->dedicated.used = heapInfo.heap_usage;
}
else
{
gpu->shared.total = heapInfo.heap_size;
gpu->shared.used = heapInfo.heap_usage;
}
}
if (options->temp)
{
if (ffamdgpu_query_sensor_info(handle, AMDGPU_INFO_SENSOR_GPU_TEMP, sizeof(value), &value) >= 0)
gpu->temperature = value;
}
ffStrbufSetS(&gpu->name, ffamdgpu_get_marketing_name(handle));
struct amdgpu_gpu_info gpuInfo;
if (ffamdgpu_query_gpu_info(handle, &gpuInfo) >= 0)
{
gpu->coreCount = (int32_t) gpuInfo.num_shader_engines;
gpu->frequency = (uint32_t) (gpuInfo.max_engine_clk / 1000u);
gpu->index = gpuInfo.asic_id;
}
if (ffamdgpu_query_sensor_info(handle, AMDGPU_INFO_SENSOR_GPU_LOAD, sizeof(value), &value) >= 0)
gpu->coreUsage = value;
ffamdgpu_device_deinitialize(handle);
return NULL;
#else
return "Fastfetch is compiled without libdrm support";
#endif
}
static void pciDetectAmdSpecific(const FFGPUOptions* options, FFGPUResult* gpu, FFstrbuf* pciDir, FFstrbuf* buffer)
{
// https://www.kernel.org/doc/html/v5.10/gpu/amdgpu.html#mem-info-vis-vram-total
@@ -186,6 +185,11 @@ static void pciDetectAmdSpecific(const FFGPUOptions* options, FFGPUResult* gpu,
ffStrbufAppendS(pciDir, "/mem_info_vis_vram_total");
if (ffReadFileBuffer(pciDir->chars, buffer) && (value = ffStrbufToUInt(buffer, 0)))
{
if (gpu->type == FF_GPU_TYPE_DISCRETE)
gpu->dedicated.total = value;
else
gpu->shared.total = value;
ffStrbufSubstrBefore(pciDir, pciDir->length - (uint32_t) strlen("/mem_info_vis_vram_total"));
ffStrbufAppendS(pciDir, "/mem_info_vis_vram_used");
if (ffReadFileBuffer(pciDir->chars, buffer) && (value = ffStrbufToUInt(buffer, 0)))
@@ -306,27 +310,6 @@ static const char* detectPci(const FFGPUOptions* options, FFlist* gpus, FFstrbuf
if (drmKey) ffStrbufSetF(&gpu->platformApi, "DRM (%s)", drmKey);
if (gpu->vendor.chars == FF_GPU_VENDOR_NAME_AMD)
{
ffStrbufAppendS(deviceDir, "/revision");
if (ffReadFileBuffer(deviceDir->chars, buffer))
{
char* pend;
uint64_t revision = strtoul(buffer->chars, &pend, 16);
if (pend != buffer->chars)
{
char query[32];
snprintf(query, sizeof(query), "%X,\t%X,", (unsigned) deviceId, (unsigned) revision);
#ifdef FF_CUSTOM_AMDGPU_IDS_PATH
ffParsePropFile(FF_STR(FF_CUSTOM_AMDGPU_IDS_PATH), query, &gpu->name);
#else
ffParsePropFileData("libdrm/amdgpu.ids", query, &gpu->name);
#endif
}
}
ffStrbufSubstrBefore(deviceDir, drmDirPathLength);
}
FF_STRBUF_AUTO_DESTROY coreName = ffStrbufCreate();
if (gpu->name.length == 0)
{
@@ -339,13 +322,38 @@ static const char* detectPci(const FFGPUOptions* options, FFlist* gpus, FFstrbuf
ffGPUParsePciIds(&pciids, subclassId, (uint16_t) vendorId, (uint16_t) deviceId, gpu, &coreName);
}
pciDetectDriver(gpu, deviceDir, buffer, drmKey);
pciDetectDriver(&gpu->driver, deviceDir, buffer, drmKey);
ffStrbufSubstrBefore(deviceDir, drmDirPathLength);
if (gpu->vendor.chars == FF_GPU_VENDOR_NAME_AMD)
{
pciDetectAmdSpecific(options, gpu, deviceDir, buffer);
ffStrbufSubstrBefore(deviceDir, drmDirPathLength);
bool ok = false;
if (drmKey && options->driverSpecific)
ok = drmDetectAmdSpecific(options, gpu, drmKey, buffer) == NULL;
if (!ok)
{
pciDetectAmdSpecific(options, gpu, deviceDir, buffer);
ffStrbufSubstrBefore(deviceDir, drmDirPathLength);
ffStrbufAppendS(deviceDir, "/revision");
if (ffReadFileBuffer(deviceDir->chars, buffer))
{
char* pend;
uint64_t revision = strtoul(buffer->chars, &pend, 16);
if (pend != buffer->chars)
{
char query[32];
snprintf(query, sizeof(query), "%X,\t%X,", (unsigned) deviceId, (unsigned) revision);
#ifdef FF_CUSTOM_AMDGPU_IDS_PATH
ffParsePropFile(FF_STR(FF_CUSTOM_AMDGPU_IDS_PATH), query, &gpu->name);
#else
ffParsePropFileData("libdrm/amdgpu.ids", query, &gpu->name);
#endif
}
}
ffStrbufSubstrBefore(deviceDir, drmDirPathLength);
}
}
else if (gpu->vendor.chars == FF_GPU_VENDOR_NAME_INTEL)
{
@@ -407,10 +415,9 @@ FF_MAYBE_UNUSED static const char* detectAsahi(FFlist* gpus, FFstrbuf* buffer, F
index += (uint32_t) strlen("apple,agx-t");
FFGPUResult* gpu = (FFGPUResult*)ffListAdd(gpus);
gpu->index = FF_GPU_INDEX_UNSET;
gpu->deviceId = strtoul(buffer->chars + index, NULL, 10);
ffStrbufInitStatic(&gpu->name, ffCPUAppleCodeToName((uint32_t) gpu->deviceId));
if (!gpu->name.length)
ffStrbufAppendF(&gpu->name, "Apple Silicon T%u", (uint32_t) gpu->deviceId);
ffStrbufInitStatic(&gpu->vendor, FF_GPU_VENDOR_NAME_APPLE);
ffStrbufInit(&gpu->driver);
ffStrbufInitF(&gpu->platformApi, "DRM (%s)", drmKey);
@@ -421,8 +428,6 @@ FF_MAYBE_UNUSED static const char* detectAsahi(FFlist* gpus, FFstrbuf* buffer, F
gpu->dedicated.total = gpu->dedicated.used = gpu->shared.total = gpu->shared.used = FF_GPU_VMEM_SIZE_UNSET;
gpu->frequency = FF_GPU_FREQUENCY_UNSET;
pciDetectDriver(gpu, drmDir, buffer, drmKey);
#if FF_HAVE_DRM
ffStrbufSetS(buffer, "/dev/dri/");
ffStrbufAppendS(buffer, drmKey);
@@ -430,12 +435,14 @@ FF_MAYBE_UNUSED static const char* detectAsahi(FFlist* gpus, FFstrbuf* buffer, F
if (fd >= 0)
{
struct drm_asahi_params_global paramsGlobal = {};
if (ioctl(fd, DRM_IOCTL_ASAHI_GET_PARAMS, &(struct drm_asahi_get_params){
if (ioctl(fd, DRM_IOCTL_ASAHI_GET_PARAMS, &(struct drm_asahi_get_params) {
.param_group = DRM_ASAHI_GET_PARAMS,
.pointer = (uint64_t) &paramsGlobal,
.size = sizeof(paramsGlobal),
}) >= 0)
{
ffStrbufSetF(&gpu->driver, "asahi %u", paramsGlobal.unstable_uabi_version);
// FIXME: They will introduce ABI breaking changes. Always check the latest version
// https://www.reddit.com/r/AsahiLinux/comments/1ei2qiv/comment/lgm0v5s/
if (paramsGlobal.unstable_uabi_version == DRM_ASAHI_UNSTABLE_UABI_VERSION)
@@ -443,9 +450,39 @@ FF_MAYBE_UNUSED static const char* detectAsahi(FFlist* gpus, FFstrbuf* buffer, F
gpu->coreCount = (int) paramsGlobal.num_cores_total_active;
gpu->frequency = paramsGlobal.max_frequency_khz / 1000;
gpu->deviceId = paramsGlobal.chip_id;
if (!gpu->name.length)
{
const char* variant = " Unknown";
switch (paramsGlobal.gpu_variant) {
case 'G':
variant = "";
break;
case 'S':
variant = " Pro";
break;
case 'C':
variant = " Max";
break;
case 'D':
variant = " Ultra";
break;
}
ffStrbufSetF(&gpu->name, "Apple M%d%s (G%d%c %02X)",
paramsGlobal.gpu_generation - 12, variant,
paramsGlobal.gpu_generation, paramsGlobal.gpu_variant,
paramsGlobal.gpu_revision + 0xA0);
}
}
}
}
if (!gpu->driver.length)
{
pciDetectDriver(&gpu->driver, drmDir, buffer, drmKey);
if (!gpu->name.length)
ffStrbufSetF(&gpu->name, "Apple Silicon T%u", (uint32_t) gpu->deviceId);
}
#endif
return NULL;