[PATCH 0/2] [GIT PULL][v2.6.33] ring-buffer: resizing issues

[PATCH 0/2] [GIT PULL][v2.6.33] ring-buffer: resizing issues

[Steven Rostedt]

Ingo,

Please pull the latest tip/tracing/urgent tree, which can be found at:

  git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-2.6-trace.git
tip/tracing/urgent


Steven Rostedt (2):
      ring-buffer: Use sync sched protection on ring buffer resizing
      ring-buffer: Move resize integrity check under reader lock

----
 kernel/trace/ring_buffer.c |   29 +++++++++++------------------
 1 files changed, 11 insertions(+), 18 deletions(-)

[PATCH 1/2] [PATCH 1/2] ring-buffer: Use sync sched protection on ring buffer resizing

[Steven Rostedt]

[-- Attachment #1: 0001-ring-buffer-Use-sync-sched-protection-on-ring-buffer.patch --]
[-- Type: text/plain, Size: 3210 bytes --]

From: Steven Rostedt <srostedt@redhat.com>

There was a comment in the ring buffer code that says the calling
layers should prevent tracing or reading of the ring buffer while
resizing. I have discovered that the tracers do not honor this
arrangement.

This patch moves the disabling and synchronizing the ring buffer to
a higher layer during resizing. This guarantees that no writes
are occurring while the resize takes place.

Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
 kernel/trace/ring_buffer.c |   25 +++++++++----------------
 1 files changed, 9 insertions(+), 16 deletions(-)

diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c
index a1ca495..0d64c51 100644
--- a/kernel/trace/ring_buffer.c
+++ b/kernel/trace/ring_buffer.c
@@ -1193,9 +1193,6 @@ rb_remove_pages(struct ring_buffer_per_cpu *cpu_buffer, unsigned nr_pages)
 	struct list_head *p;
 	unsigned i;
 
-	atomic_inc(&cpu_buffer->record_disabled);
-	synchronize_sched();
-
 	spin_lock_irq(&cpu_buffer->reader_lock);
 	rb_head_page_deactivate(cpu_buffer);
 
@@ -1214,9 +1211,6 @@ rb_remove_pages(struct ring_buffer_per_cpu *cpu_buffer, unsigned nr_pages)
 	spin_unlock_irq(&cpu_buffer->reader_lock);
 
 	rb_check_pages(cpu_buffer);
-
-	atomic_dec(&cpu_buffer->record_disabled);
-
 }
 
 static void
@@ -1227,9 +1221,6 @@ rb_insert_pages(struct ring_buffer_per_cpu *cpu_buffer,
 	struct list_head *p;
 	unsigned i;
 
-	atomic_inc(&cpu_buffer->record_disabled);
-	synchronize_sched();
-
 	spin_lock_irq(&cpu_buffer->reader_lock);
 	rb_head_page_deactivate(cpu_buffer);
 
@@ -1245,8 +1236,6 @@ rb_insert_pages(struct ring_buffer_per_cpu *cpu_buffer,
 	spin_unlock_irq(&cpu_buffer->reader_lock);
 
 	rb_check_pages(cpu_buffer);
-
-	atomic_dec(&cpu_buffer->record_disabled);
 }
 
 /**
@@ -1254,11 +1243,6 @@ rb_insert_pages(struct ring_buffer_per_cpu *cpu_buffer,
  * @buffer: the buffer to resize.
  * @size: the new size.
  *
- * The tracer is responsible for making sure that the buffer is
- * not being used while changing the size.
- * Note: We may be able to change the above requirement by using
- *  RCU synchronizations.
- *
  * Minimum size is 2 * BUF_PAGE_SIZE.
  *
  * Returns -1 on failure.
@@ -1290,6 +1274,11 @@ int ring_buffer_resize(struct ring_buffer *buffer, unsigned long size)
 	if (size == buffer_size)
 		return size;
 
+	atomic_inc(&buffer->record_disabled);
+
+	/* Make sure all writers are done with this buffer. */
+	synchronize_sched();
+
 	mutex_lock(&buffer->mutex);
 	get_online_cpus();
 
@@ -1352,6 +1341,8 @@ int ring_buffer_resize(struct ring_buffer *buffer, unsigned long size)
 	put_online_cpus();
 	mutex_unlock(&buffer->mutex);
 
+	atomic_dec(&buffer->record_disabled);
+
 	return size;
 
  free_pages:
@@ -1361,6 +1352,7 @@ int ring_buffer_resize(struct ring_buffer *buffer, unsigned long size)
 	}
 	put_online_cpus();
 	mutex_unlock(&buffer->mutex);
+	atomic_dec(&buffer->record_disabled);
 	return -ENOMEM;
 
 	/*
@@ -1370,6 +1362,7 @@ int ring_buffer_resize(struct ring_buffer *buffer, unsigned long size)
  out_fail:
 	put_online_cpus();
 	mutex_unlock(&buffer->mutex);
+	atomic_dec(&buffer->record_disabled);
 	return -1;
 }
 EXPORT_SYMBOL_GPL(ring_buffer_resize);
-- 
1.6.5

[PATCH 2/2] [PATCH 2/2] ring-buffer: Move resize integrity check under reader lock

[Steven Rostedt]

[-- Attachment #1: 0002-ring-buffer-Move-resize-integrity-check-under-reader.patch --]
[-- Type: text/plain, Size: 1454 bytes --]

From: Steven Rostedt <srostedt@redhat.com>

While using an application that does splice on the ftrace ring
buffer at start up, I triggered an integrity check failure.

Looking into this, I discovered that resizing the buffer performs
an integrity check after the buffer is resized. This check unfortunately
is preformed after it releases the reader lock. If a reader is
reading the buffer it may cause the integrity check to trigger a
false failure.

This patch simply moves the integrity checker under the protection
of the ring buffer reader lock.

Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
 kernel/trace/ring_buffer.c |    8 ++++----
 1 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c
index 0d64c51..eccb4cf 100644
--- a/kernel/trace/ring_buffer.c
+++ b/kernel/trace/ring_buffer.c
@@ -1208,9 +1208,9 @@ rb_remove_pages(struct ring_buffer_per_cpu *cpu_buffer, unsigned nr_pages)
 		return;
 
 	rb_reset_cpu(cpu_buffer);
-	spin_unlock_irq(&cpu_buffer->reader_lock);
-
 	rb_check_pages(cpu_buffer);
+
+	spin_unlock_irq(&cpu_buffer->reader_lock);
 }
 
 static void
@@ -1233,9 +1233,9 @@ rb_insert_pages(struct ring_buffer_per_cpu *cpu_buffer,
 		list_add_tail(&bpage->list, cpu_buffer->pages);
 	}
 	rb_reset_cpu(cpu_buffer);
-	spin_unlock_irq(&cpu_buffer->reader_lock);
-
 	rb_check_pages(cpu_buffer);
+
+	spin_unlock_irq(&cpu_buffer->reader_lock);
 }
 
 /**
-- 
1.6.5

Re: [PATCH 0/2] [GIT PULL][v2.6.33] ring-buffer: resizing issues

[Ingo Molnar]

* Steven Rostedt <rostedt@goodmis.org> wrote:

> 
> Ingo,
> 
> Please pull the latest tip/tracing/urgent tree, which can be found at:
> 
>   git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-2.6-trace.git
> tip/tracing/urgent
> 
> 
> Steven Rostedt (2):
>       ring-buffer: Use sync sched protection on ring buffer resizing
>       ring-buffer: Move resize integrity check under reader lock
> 
> ----
>  kernel/trace/ring_buffer.c |   29 +++++++++++------------------
>  1 files changed, 11 insertions(+), 18 deletions(-)

Pulled, thanks Steve!

	Ingo