From: James Zhu <James.Zhu@amd.com>
To: <amd-gfx@lists.freedesktop.org>, <alexander.deucher@amd.com>,
<Bing.Ma@amd.com>, <David.Francis@amd.com>
Cc: <Jesse.Zhang@amd.com>, <Jenny-Jing.Liu@amd.com>, <jamesz@amd.com>
Subject: [PATCH 1/2] amdgpu: Add profiler IOCTL interface for performance monitoring
Date: Mon, 13 Apr 2026 15:29:46 -0400 [thread overview]
Message-ID: <20260413192947.3145878-1-James.Zhu@amd.com> (raw)
This patch introduces the foundational profiler infrastructure for AMD GPUs,
enabling userspace access to performance monitoring capabilities including:
- Performance Monitoring Counters (PMC)
- Performance Counter Sampling (PC Sampling)
- Streaming Performance Monitor (SPM)
The implementation includes:
- New DRM_AMDGPU_PROFILER IOCTL interface with version query support
- amdgpu_profiler_version() wrapper function for userspace
- Profiler operation enumeration and argument structures in the kernel API
- Build system integration for the new profiler module
The version query operation allows userspace to determine profiler capability
and compatibility before attempting to use advanced profiling features. Future
patches will extend this with additional profiler operations (acquire, release,
configure sampling buffers, etc.).
This foundation enables profiling tools and performance analysis frameworks to
access GPU performance data directly from userspace, supporting performance
debugging and optimization workflows.
Signed-off-by: James Zhu <James.Zhu@amd.com>
---
amdgpu/amdgpu.h | 9 ++++++++
amdgpu/amdgpu_profiler.c | 46 ++++++++++++++++++++++++++++++++++++++++
amdgpu/meson.build | 2 +-
include/drm/amdgpu_drm.h | 19 +++++++++++++++++
4 files changed, 75 insertions(+), 1 deletion(-)
create mode 100644 amdgpu/amdgpu_profiler.c
diff --git a/amdgpu/amdgpu.h b/amdgpu/amdgpu.h
index 53144f59..4ec1f6b6 100644
--- a/amdgpu/amdgpu.h
+++ b/amdgpu/amdgpu.h
@@ -2120,6 +2120,15 @@ int amdgpu_userq_wait(amdgpu_device_handle dev,
int amdgpu_cwsr_set_l2_trap_handler(amdgpu_device_handle dev,
uint64_t tba_addr, uint64_t tba_size,
uint64_t tma_addr, uint64_t tma_size);
+
+/**
+ * Acquire profiler version
+ * \param dev - \c [in] device handle
+ *
+ * \return 0 on success otherwise POSIX Error code
+ */
+int amdgpu_profiler_version(amdgpu_device_handle dev);
+
#ifdef __cplusplus
}
#endif
diff --git a/amdgpu/amdgpu_profiler.c b/amdgpu/amdgpu_profiler.c
new file mode 100644
index 00000000..8d4dffe4
--- /dev/null
+++ b/amdgpu/amdgpu_profiler.c
@@ -0,0 +1,46 @@
+/*
+ * Copyright 2026 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.
+ *
+ */
+
+#include <string.h>
+#include <errno.h>
+#include "xf86drm.h"
+#include "amdgpu_drm.h"
+#include "amdgpu_internal.h"
+
+drm_public int
+amdgpu_profiler_version(amdgpu_device_handle dev)
+{
+ int ret;
+ struct drm_amdgpu_profiler_args user_arg;
+
+ if (!dev)
+ return -EINVAL;
+
+ memset(&user_arg, 0, sizeof(user_arg));
+ user_arg.op = AMDGPU_PROFILER_VERSION;
+
+ ret = drmCommandWriteRead(dev->fd, DRM_AMDGPU_PROFILER,
+ &user_arg, sizeof(user_arg));
+
+ return ret;
+}
diff --git a/amdgpu/meson.build b/amdgpu/meson.build
index 3962d32c..d781f2e9 100644
--- a/amdgpu/meson.build
+++ b/amdgpu/meson.build
@@ -27,7 +27,7 @@ libdrm_amdgpu = library(
files(
'amdgpu_asic_id.c', 'amdgpu_bo.c', 'amdgpu_cs.c', 'amdgpu_device.c',
'amdgpu_gpu_info.c', 'amdgpu_vamgr.c', 'amdgpu_vm.c', 'handle_table.c',
- 'amdgpu_userq.c',
+ 'amdgpu_userq.c', 'amdgpu_profiler.c',
),
config_file,
],
diff --git a/include/drm/amdgpu_drm.h b/include/drm/amdgpu_drm.h
index ef12e725..307242ac 100644
--- a/include/drm/amdgpu_drm.h
+++ b/include/drm/amdgpu_drm.h
@@ -58,6 +58,7 @@ extern "C" {
#define DRM_AMDGPU_USERQ_SIGNAL 0x17
#define DRM_AMDGPU_USERQ_WAIT 0x18
#define DRM_AMDGPU_CWSR 0x20
+#define DRM_AMDGPU_PROFILER 0x21
#define DRM_IOCTL_AMDGPU_GEM_CREATE DRM_IOWR(DRM_COMMAND_BASE + DRM_AMDGPU_GEM_CREATE, union drm_amdgpu_gem_create)
#define DRM_IOCTL_AMDGPU_GEM_MMAP DRM_IOWR(DRM_COMMAND_BASE + DRM_AMDGPU_GEM_MMAP, union drm_amdgpu_gem_mmap)
@@ -79,6 +80,7 @@ extern "C" {
#define DRM_IOCTL_AMDGPU_USERQ_SIGNAL DRM_IOWR(DRM_COMMAND_BASE + DRM_AMDGPU_USERQ_SIGNAL, struct drm_amdgpu_userq_signal)
#define DRM_IOCTL_AMDGPU_USERQ_WAIT DRM_IOWR(DRM_COMMAND_BASE + DRM_AMDGPU_USERQ_WAIT, struct drm_amdgpu_userq_wait)
#define DRM_IOCTL_AMDGPU_CWSR DRM_IOWR(DRM_COMMAND_BASE + DRM_AMDGPU_CWSR, union drm_amdgpu_cwsr)
+#define DRM_IOCTL_AMDGPU_PROFILER DRM_IOWR(DRM_COMMAND_BASE + DRM_AMDGPU_PROFILER, struct drm_amdgpu_profiler_args)
/**
* DOC: memory domains
@@ -1695,6 +1697,23 @@ struct drm_amdgpu_info_gpuvm_fault {
#define AMDGPU_FAMILY_GC_11_5_0 150 /* GC 11.5.0 */
#define AMDGPU_FAMILY_GC_12_0_0 152 /* GC 12.0.0 */
+/*
+ * Supported Profiler Operations
+ */
+enum drm_amdgpu_profiler_ops {
+ AMDGPU_PROFILER_VERSION = 0,
+};
+
+struct drm_amdgpu_profiler_args {
+ __u32 op; /* amdgpu_profiler_op */
+ union {
+ __u32 version; /* AMDGPU_PROFILER_VERSION_NUM
+ * lower 16 bit: minor
+ * higher 16 bit: major
+ */
+ };
+};
+
#if defined(__cplusplus)
}
#endif
--
2.34.1
next reply other threads:[~2026-04-13 19:30 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-04-13 19:29 James Zhu [this message]
2026-04-13 19:29 ` [PATCH 2/2] amdgpu: Add Streaming Performance Monitor (SPM) data collection interface James Zhu
2026-05-13 23:56 ` Ma, Bing
2026-05-15 17:14 ` [PATCH v2 " James Zhu
2026-04-16 14:06 ` [PATCH 1/2] amdgpu: Add profiler IOCTL interface for performance monitoring James Zhu
2026-05-13 23:55 ` Ma, Bing
2026-05-15 15:34 ` James Zhu
2026-05-15 17:13 ` [PATCH v2 " James Zhu
2026-05-15 17:17 ` James Zhu
2026-05-15 17:20 ` [PATCH v2 1/2] amdgpu: " James Zhu
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20260413192947.3145878-1-James.Zhu@amd.com \
--to=james.zhu@amd.com \
--cc=Bing.Ma@amd.com \
--cc=David.Francis@amd.com \
--cc=Jenny-Jing.Liu@amd.com \
--cc=Jesse.Zhang@amd.com \
--cc=alexander.deucher@amd.com \
--cc=amd-gfx@lists.freedesktop.org \
--cc=jamesz@amd.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.