Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Tue, 13 Jun 2023 09:35:13 +0100
In-Reply-To: <20230613083513.3312612-1-vdonnefort@google.com>
Mime-Version: 1.0
References: <20230613083513.3312612-1-vdonnefort@google.com>
Message-ID: <20230613083513.3312612-3-vdonnefort@google.com>
Subject: [PATCH v4 2/2] tracing: Allow user-space mapping of the ring-buffer
From:   Vincent Donnefort <vdonnefort@google.com>
To:     rostedt@goodmis.org, mhiramat@kernel.org,
        linux-kernel@vger.kernel.org, linux-trace-kernel@vger.kernel.org
Cc:     kernel-team@android.com, Vincent Donnefort <vdonnefort@google.com>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

Currently, user-space extracts data from the ring-buffer via splice,
which is handy for storage or network sharing. However, due to splice
limitations, it is imposible to do real-time analysis without a copy.

A solution for that problem is to let the user-space map the ring-buffer
directly.

The mapping exposed via the per-CPU file trace_pipe_raw. The first page
is the meta-page and is followed by each page of the ring-buffer,
ordered by their unique page ID. It is therefore easy to translate a
page-ID to an offset in the mapping.

  * Meta-page -- include/uapi/linux/trace_mmap.h for a description
  * Page ID 0
  * Page ID 1
     ...

The mapper must then do what use to be the kernel job: swap the reader
with the head. This is done with a newly introduced ioctl:
TRACE_MMAP_IOCTL_GET_READER_PAGE.

Signed-off-by: Vincent Donnefort <vdonnefort@google.com>

diff --git a/include/uapi/linux/trace_mmap.h b/include/uapi/linux/trace_mmap.h
index 653176cc50bc..fd323a92cf78 100644
--- a/include/uapi/linux/trace_mmap.h
+++ b/include/uapi/linux/trace_mmap.h
@@ -23,4 +23,6 @@ struct ring_buffer_meta {
 	} reader_page;
 };
 
+#define TRACE_MMAP_IOCTL_GET_READER_PAGE	_IO('T', 0x1)
+
 #endif /* _UAPI_TRACE_MMAP_H_ */
diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index 64a4dde073ef..ccc0e1dd7d0d 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -6432,7 +6432,7 @@ static void tracing_set_nop(struct trace_array *tr)
 {
 	if (tr->current_trace == &nop_trace)
 		return;
-	
+
 	tr->current_trace->enabled--;
 
 	if (tr->current_trace->reset)
@@ -8465,15 +8465,21 @@ tracing_buffers_splice_read(struct file *file, loff_t *ppos,
 	return ret;
 }
 
-/* An ioctl call with cmd 0 to the ring buffer file will wake up all waiters */
 static long tracing_buffers_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
 {
 	struct ftrace_buffer_info *info = file->private_data;
 	struct trace_iterator *iter = &info->iter;
 
-	if (cmd)
-		return -ENOIOCTLCMD;
+	if (cmd == TRACE_MMAP_IOCTL_GET_READER_PAGE)
+		return ring_buffer_map_get_reader_page(iter->array_buffer->buffer,
+						       iter->cpu_file);
+	else if (cmd)
+		return -ENOTTY;
 
+	/*
+	 * An ioctl call with cmd 0 to the ring buffer file will wake up all
+	 * waiters
+	 */
 	mutex_lock(&trace_types_lock);
 
 	iter->wait_index++;
@@ -8486,6 +8492,63 @@ static long tracing_buffers_ioctl(struct file *file, unsigned int cmd, unsigned
 	return 0;
 }
 
+static vm_fault_t tracing_buffers_mmap_fault(struct vm_fault *vmf)
+{
+	struct ftrace_buffer_info *info = vmf->vma->vm_file->private_data;
+	struct trace_iterator *iter = &info->iter;
+	vm_fault_t ret = VM_FAULT_SIGBUS;
+	struct page *page;
+
+	page = ring_buffer_map_fault(iter->array_buffer->buffer, iter->cpu_file,
+				     vmf->pgoff);
+	if (!page)
+		return ret;
+
+	vmf->page = page;
+
+	get_page(vmf->page);
+	vmf->page->mapping = vmf->vma->vm_file->f_mapping;
+	vmf->page->index   = vmf->pgoff;
+
+	return 0;
+}
+
+static void tracing_buffers_mmap_close(struct vm_area_struct *vma)
+{
+	struct ftrace_buffer_info *info = vma->vm_file->private_data;
+	struct trace_iterator *iter = &info->iter;
+
+	ring_buffer_unmap(iter->array_buffer->buffer, iter->cpu_file);
+}
+
+static void tracing_buffers_mmap_open(struct vm_area_struct *vma)
+{
+	struct ftrace_buffer_info *info = vma->vm_file->private_data;
+	struct trace_iterator *iter = &info->iter;
+
+	WARN_ON(ring_buffer_map(iter->array_buffer->buffer, iter->cpu_file));
+}
+
+static const struct vm_operations_struct tracing_buffers_vmops = {
+	.open		= tracing_buffers_mmap_open,
+	.close		= tracing_buffers_mmap_close,
+	.fault		= tracing_buffers_mmap_fault,
+};
+
+static int tracing_buffers_mmap(struct file *filp, struct vm_area_struct *vma)
+{
+	struct ftrace_buffer_info *info = filp->private_data;
+	struct trace_iterator *iter = &info->iter;
+
+	if (vma->vm_flags & VM_WRITE)
+		return -EPERM;
+
+	vm_flags_mod(vma, VM_DONTCOPY | VM_DONTDUMP, VM_MAYWRITE);
+	vma->vm_ops = &tracing_buffers_vmops;
+
+	return ring_buffer_map(iter->array_buffer->buffer, iter->cpu_file);
+}
+
 static const struct file_operations tracing_buffers_fops = {
 	.open		= tracing_buffers_open,
 	.read		= tracing_buffers_read,
@@ -8494,6 +8557,7 @@ static const struct file_operations tracing_buffers_fops = {
 	.splice_read	= tracing_buffers_splice_read,
 	.unlocked_ioctl = tracing_buffers_ioctl,
 	.llseek		= no_llseek,
+	.mmap		= tracing_buffers_mmap,
 };
 
 static ssize_t
-- 
2.41.0.162.gfafddb0af9-goog