Thread (2 messages) 2 messages, 2 authors, 2025-12-26

Re: [PATCH v11] dma-buf: add some tracepoints to debug.

From: Steven Rostedt <rostedt@goodmis.org>
Date: 2025-12-26 16:42:26
Also in: dri-devel, linux-media, lkml

On Thu, 25 Dec 2025 20:11:19 +0800
Xiang Gao [off-list ref] wrote:
From: gaoxiang17 <redacted>

Since we can only inspect dmabuf by iterating over process FDs or the
dmabuf_list, we need to add our own tracepoints to track its status in
real time in production.

For example:
   binder:3016_1-3102    [006] ...1.   255.126521: dma_buf_export: exp_name=qcom,system size=12685312 ino=2738
   binder:3016_1-3102    [006] ...1.   255.126528: dma_buf_fd: exp_name=qcom,system size=12685312 ino=2738 fd=8
   binder:3016_1-3102    [006] ...1.   255.126642: dma_buf_mmap_internal: exp_name=qcom,system size=28672 ino=2739
     kworker/6:1-86      [006] ...1.   255.127194: dma_buf_put: exp_name=qcom,system size=12685312 ino=2738
    RenderThread-9293    [006] ...1.   316.618179: dma_buf_get: exp_name=qcom,system size=12771328 ino=2762 fd=176
    RenderThread-9293    [006] ...1.   316.618195: dma_buf_dynamic_attach: exp_name=qcom,system size=12771328 ino=2762 attachment:ffffff880a18dd00 is_dynamic=0 dev_name=kgsl-3d0
    RenderThread-9293    [006] ...1.   318.878220: dma_buf_detach: exp_name=qcom,system size=12771328 ino=2762 attachment:ffffff880a18dd00 is_dynamic=0 dev_name=kgsl-3d0

Signed-off-by: Xiang Gao <redacted>
---

Changes since v10: https://lore.kernel.org/all/20251224013455.1649879-1-gxxa03070307@gmail.com/ (local)
You only need to show the last revision (but since you haven't showed the
others, before it's good to show them now). But you also want to say "what changed"

For example:

- Used DEFINE_EVENT_CONDITION() to move the condition branch into the tracepoint.

Changes since v9: https://lore.kernel.org/all/20251223032749.1371913-1-gxxa03070307@gmail.com/ (local)
Changes since v8: https://lore.kernel.org/all/20251218062853.819744-1-gxxa03070307@gmail.com/ (local)
Changes since v7: https://lore.kernel.org/all/20251217105132.643300-1-gxxa03070307@gmail.com/ (local)
Changes since v6: https://lore.kernel.org/all/20251216063952.516364-1-gxxa03070307@gmail.com/ (local)
Changes since v5: https://lore.kernel.org/all/20251201112148.843572-1-gxxa03070307@gmail.com/ (local)
Changes since v4: https://lore.kernel.org/all/20251128085215.550100-1-gxxa03070307@gmail.com/ (local)
Changes since v3: https://lore.kernel.org/all/20251127004352.376307-1-gxxa03070307@gmail.com/ (local)
Changes since v2: https://lore.kernel.org/all/20251125162949.220488-1-gxxa03070307@gmail.com/ (local)
Changes since v1: https://lore.kernel.org/all/20251124133648.72668-1-gxxa03070307@gmail.com/ (local)

 drivers/dma-buf/dma-buf.c      |  48 +++++++++-
 include/trace/events/dma_buf.h | 159 +++++++++++++++++++++++++++++++++
 2 files changed, 205 insertions(+), 2 deletions(-)
 create mode 100644 include/trace/events/dma_buf.h
+/*
+ * dmabuf->name must be accessed with holding dmabuf->name_lock.
+ * we need to take the lock around the tracepoint call itself where
+ * it is called in the code.
+ *
+ * Note: FUNC##_enabled() is a static branch that will only
+ *       be set when the trace event is enabled.
+ */
+#define DMA_BUF_TRACE(FUNC, ...)					\
+	do {								\
+		/* Always expose lock if lockdep is enabled */		\
+		if (IS_ENABLED(CONFIG_LOCKDEP) || FUNC##_enabled()) {	\
+			guard(spinlock)(&dmabuf->name_lock);		\
+			FUNC(__VA_ARGS__);				\
+		}							\
+	} while (0)
+
 static inline int is_dma_buf_file(struct file *);
 
+
+DEFINE_EVENT_CONDITION(dma_buf_fd, dma_buf_fd,
+
+	TP_PROTO(struct dma_buf *dmabuf, int fd),
+
+	TP_ARGS(dmabuf, fd),
+
+	TP_CONDITION(fd >= 0)
+);
+

From a tracing point of view (It's up to the dma maintainers to decide to
take this patch):

Reviewed-by: Steven Rostedt (Google) <rostedt@goodmis.org>

-- Steve
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help