* [PATCH v2 0/2] tracing: Clean up how iter is freed
@ 2023-07-15 14:12 Steven Rostedt
2023-07-15 14:12 ` [PATCH v2 1/2] tracing: Remove unnecessary copying of tr->current_trace Steven Rostedt
2023-07-15 14:12 ` [PATCH v2 2/2] tracing: Add free_trace_iter_content() helper function Steven Rostedt
0 siblings, 2 replies; 5+ messages in thread
From: Steven Rostedt @ 2023-07-15 14:12 UTC (permalink / raw)
To: linux-kernel, linux-trace-kernel
Cc: Masami Hiramatsu, Mark Rutland, Andrew Morton
The trace iterator is used in various interfaces and needs to be consistent
in how it is cleaned up. Add a helper function to clean up its content. But
before doing so, I noticed that iter->trace is allocated then the content
of tr->current_trace is copied to it. There's no reason for this, so the
first patch removes that allocation and just points to the content of
tr->current_trace, as tr->current_trace can change, but the content should
not.
Changes since v1: https://lore.kernel.org/linux-trace-kernel/20230713114510.04c452ca@gandalf.local.home/
- Updated the change log of patch 1, as I remembered why the copy was done.
- Remove the allocation and copy of the second patch. I started with the
second patch and noticed that there was an inconsistency with the
iter->trace. Where in one place it was allocated and copied and the second
place it was just pointing to the tr->current_trace. I first started to make it
be a copy as well, but then realized that the copy wasn't needed. I created
the first patch but forgot to remove the copy and it ended up in the second
patch.
Steven Rostedt (Google) (2):
tracing: Remove unnecessary copying of tr->current_trace
tracing: Add free_trace_iter_content() helper function
----
kernel/trace/trace.c | 55 ++++++++++++++++++++++++----------------------------
1 file changed, 25 insertions(+), 30 deletions(-)
^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH v2 1/2] tracing: Remove unnecessary copying of tr->current_trace
2023-07-15 14:12 [PATCH v2 0/2] tracing: Clean up how iter is freed Steven Rostedt
@ 2023-07-15 14:12 ` Steven Rostedt
2023-07-15 14:12 ` [PATCH v2 2/2] tracing: Add free_trace_iter_content() helper function Steven Rostedt
1 sibling, 0 replies; 5+ messages in thread
From: Steven Rostedt @ 2023-07-15 14:12 UTC (permalink / raw)
To: linux-kernel, linux-trace-kernel
Cc: Masami Hiramatsu, Mark Rutland, Andrew Morton
From: "Steven Rostedt (Google)" <rostedt@goodmis.org>
The iterator allocated a descriptor to copy the current_trace. This was done
with the assumption that the function pointers might change. But this was a
false assuption, as it does not change. There's no reason to make a copy of the
current_trace and just use the pointer it points to. This removes needing to
manage freeing the descriptor. Worse yet, there's locations that the iterator
is used but does make a copy and just uses the pointer. This could cause the
actual pointer to the trace descriptor to be freed and not the allocated copy.
This is more of a clean up than a fix.
Fixes: d7350c3f45694 ("tracing/core: make the read callbacks reentrants")
Reviewed-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
---
kernel/trace/trace.c | 22 +++-------------------
1 file changed, 3 insertions(+), 19 deletions(-)
diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index be847d45d81c..1c370ffbe062 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -4205,15 +4205,9 @@ static void *s_start(struct seq_file *m, loff_t *pos)
loff_t l = 0;
int cpu;
- /*
- * copy the tracer to avoid using a global lock all around.
- * iter->trace is a copy of current_trace, the pointer to the
- * name may be used instead of a strcmp(), as iter->trace->name
- * will point to the same string as current_trace->name.
- */
mutex_lock(&trace_types_lock);
- if (unlikely(tr->current_trace && iter->trace->name != tr->current_trace->name))
- *iter->trace = *tr->current_trace;
+ if (unlikely(tr->current_trace != iter->trace))
+ iter->trace = tr->current_trace;
mutex_unlock(&trace_types_lock);
#ifdef CONFIG_TRACER_MAX_TRACE
@@ -4862,16 +4856,8 @@ __tracing_open(struct inode *inode, struct file *file, bool snapshot)
iter->fmt = NULL;
iter->fmt_size = 0;
- /*
- * We make a copy of the current tracer to avoid concurrent
- * changes on it while we are reading.
- */
mutex_lock(&trace_types_lock);
- iter->trace = kzalloc(sizeof(*iter->trace), GFP_KERNEL);
- if (!iter->trace)
- goto fail;
-
- *iter->trace = *tr->current_trace;
+ iter->trace = tr->current_trace;
if (!zalloc_cpumask_var(&iter->started, GFP_KERNEL))
goto fail;
@@ -4936,7 +4922,6 @@ __tracing_open(struct inode *inode, struct file *file, bool snapshot)
fail:
mutex_unlock(&trace_types_lock);
- kfree(iter->trace);
kfree(iter->temp);
kfree(iter->buffer_iter);
release:
@@ -5021,7 +5006,6 @@ static int tracing_release(struct inode *inode, struct file *file)
free_cpumask_var(iter->started);
kfree(iter->fmt);
kfree(iter->temp);
- kfree(iter->trace);
kfree(iter->buffer_iter);
seq_release_private(inode, file);
--
2.40.1
^ permalink raw reply related [flat|nested] 5+ messages in thread
* [PATCH v2 2/2] tracing: Add free_trace_iter_content() helper function
2023-07-15 14:12 [PATCH v2 0/2] tracing: Clean up how iter is freed Steven Rostedt
2023-07-15 14:12 ` [PATCH v2 1/2] tracing: Remove unnecessary copying of tr->current_trace Steven Rostedt
@ 2023-07-15 14:12 ` Steven Rostedt
2023-07-26 13:42 ` Masami Hiramatsu
1 sibling, 1 reply; 5+ messages in thread
From: Steven Rostedt @ 2023-07-15 14:12 UTC (permalink / raw)
To: linux-kernel, linux-trace-kernel
Cc: Masami Hiramatsu, Mark Rutland, Andrew Morton
From: "Steven Rostedt (Google)" <rostedt@goodmis.org>
As the trace iterator is created and used by various interfaces, the clean
up of it needs to be consistent. Create a free_trace_iter_content() helper
function that frees the content of the iterator and use that to clean it
up in all places that it is used.
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
---
kernel/trace/trace.c | 33 ++++++++++++++++++++++-----------
1 file changed, 22 insertions(+), 11 deletions(-)
diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index 1c370ffbe062..8775930aa545 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -4815,6 +4815,25 @@ static const struct seq_operations tracer_seq_ops = {
.show = s_show,
};
+/*
+ * Note, as iter itself can be allocated and freed in different
+ * ways, this function is only used to free its content, and not
+ * the iterator itself. The only requirement to all the allocations
+ * is that it must zero all fields (kzalloc), as freeing works with
+ * ethier allocated content or NULL.
+ */
+static void free_trace_iter_content(struct trace_iterator *iter)
+{
+ /* The fmt is either NULL, allocated or points to static_fmt_buf */
+ if (iter->fmt != static_fmt_buf)
+ kfree(iter->fmt);
+
+ kfree(iter->temp);
+ kfree(iter->buffer_iter);
+ mutex_destroy(&iter->mutex);
+ free_cpumask_var(iter->started);
+}
+
static struct trace_iterator *
__tracing_open(struct inode *inode, struct file *file, bool snapshot)
{
@@ -4922,8 +4941,7 @@ __tracing_open(struct inode *inode, struct file *file, bool snapshot)
fail:
mutex_unlock(&trace_types_lock);
- kfree(iter->temp);
- kfree(iter->buffer_iter);
+ free_trace_iter_content(iter);
release:
seq_release_private(inode, file);
return ERR_PTR(-ENOMEM);
@@ -5002,11 +5020,7 @@ static int tracing_release(struct inode *inode, struct file *file)
mutex_unlock(&trace_types_lock);
- mutex_destroy(&iter->mutex);
- free_cpumask_var(iter->started);
- kfree(iter->fmt);
- kfree(iter->temp);
- kfree(iter->buffer_iter);
+ free_trace_iter_content(iter);
seq_release_private(inode, file);
return 0;
@@ -6763,10 +6777,7 @@ static int tracing_release_pipe(struct inode *inode, struct file *file)
mutex_unlock(&trace_types_lock);
- free_cpumask_var(iter->started);
- kfree(iter->fmt);
- kfree(iter->temp);
- mutex_destroy(&iter->mutex);
+ free_trace_iter_content(iter);
kfree(iter);
trace_array_put(tr);
--
2.40.1
^ permalink raw reply related [flat|nested] 5+ messages in thread
* Re: [PATCH v2 2/2] tracing: Add free_trace_iter_content() helper function
2023-07-15 14:12 ` [PATCH v2 2/2] tracing: Add free_trace_iter_content() helper function Steven Rostedt
@ 2023-07-26 13:42 ` Masami Hiramatsu
2023-07-28 23:08 ` Steven Rostedt
0 siblings, 1 reply; 5+ messages in thread
From: Masami Hiramatsu @ 2023-07-26 13:42 UTC (permalink / raw)
To: Steven Rostedt
Cc: linux-kernel, linux-trace-kernel, Masami Hiramatsu, Mark Rutland,
Andrew Morton
On Sat, 15 Jul 2023 10:12:15 -0400
Steven Rostedt <rostedt@goodmis.org> wrote:
> From: "Steven Rostedt (Google)" <rostedt@goodmis.org>
>
> As the trace iterator is created and used by various interfaces, the clean
> up of it needs to be consistent. Create a free_trace_iter_content() helper
> function that frees the content of the iterator and use that to clean it
> up in all places that it is used.
>
> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Looks good to me.
Reviewed-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
BTW, this adds iter->fmt != static_fmt_buf check. Is it a kind of fix?
Thank you,
> ---
> kernel/trace/trace.c | 33 ++++++++++++++++++++++-----------
> 1 file changed, 22 insertions(+), 11 deletions(-)
>
> diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
> index 1c370ffbe062..8775930aa545 100644
> --- a/kernel/trace/trace.c
> +++ b/kernel/trace/trace.c
> @@ -4815,6 +4815,25 @@ static const struct seq_operations tracer_seq_ops = {
> .show = s_show,
> };
>
> +/*
> + * Note, as iter itself can be allocated and freed in different
> + * ways, this function is only used to free its content, and not
> + * the iterator itself. The only requirement to all the allocations
> + * is that it must zero all fields (kzalloc), as freeing works with
> + * ethier allocated content or NULL.
> + */
> +static void free_trace_iter_content(struct trace_iterator *iter)
> +{
> + /* The fmt is either NULL, allocated or points to static_fmt_buf */
> + if (iter->fmt != static_fmt_buf)
> + kfree(iter->fmt);
> +
> + kfree(iter->temp);
> + kfree(iter->buffer_iter);
> + mutex_destroy(&iter->mutex);
> + free_cpumask_var(iter->started);
> +}
> +
> static struct trace_iterator *
> __tracing_open(struct inode *inode, struct file *file, bool snapshot)
> {
> @@ -4922,8 +4941,7 @@ __tracing_open(struct inode *inode, struct file *file, bool snapshot)
>
> fail:
> mutex_unlock(&trace_types_lock);
> - kfree(iter->temp);
> - kfree(iter->buffer_iter);
> + free_trace_iter_content(iter);
> release:
> seq_release_private(inode, file);
> return ERR_PTR(-ENOMEM);
> @@ -5002,11 +5020,7 @@ static int tracing_release(struct inode *inode, struct file *file)
>
> mutex_unlock(&trace_types_lock);
>
> - mutex_destroy(&iter->mutex);
> - free_cpumask_var(iter->started);
> - kfree(iter->fmt);
> - kfree(iter->temp);
> - kfree(iter->buffer_iter);
> + free_trace_iter_content(iter);
> seq_release_private(inode, file);
>
> return 0;
> @@ -6763,10 +6777,7 @@ static int tracing_release_pipe(struct inode *inode, struct file *file)
>
> mutex_unlock(&trace_types_lock);
>
> - free_cpumask_var(iter->started);
> - kfree(iter->fmt);
> - kfree(iter->temp);
> - mutex_destroy(&iter->mutex);
> + free_trace_iter_content(iter);
> kfree(iter);
>
> trace_array_put(tr);
> --
> 2.40.1
--
Masami Hiramatsu (Google) <mhiramat@kernel.org>
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH v2 2/2] tracing: Add free_trace_iter_content() helper function
2023-07-26 13:42 ` Masami Hiramatsu
@ 2023-07-28 23:08 ` Steven Rostedt
0 siblings, 0 replies; 5+ messages in thread
From: Steven Rostedt @ 2023-07-28 23:08 UTC (permalink / raw)
To: Masami Hiramatsu (Google)
Cc: linux-kernel, linux-trace-kernel, Mark Rutland, Andrew Morton
On Wed, 26 Jul 2023 22:42:13 +0900
Masami Hiramatsu (Google) <mhiramat@kernel.org> wrote:
> On Sat, 15 Jul 2023 10:12:15 -0400
> Steven Rostedt <rostedt@goodmis.org> wrote:
>
> > From: "Steven Rostedt (Google)" <rostedt@goodmis.org>
> >
> > As the trace iterator is created and used by various interfaces, the clean
> > up of it needs to be consistent. Create a free_trace_iter_content() helper
> > function that frees the content of the iterator and use that to clean it
> > up in all places that it is used.
> >
> > Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
>
> Looks good to me.
>
> Reviewed-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
>
> BTW, this adds iter->fmt != static_fmt_buf check. Is it a kind of fix?
No, because all of the callers shouldn't actually set it to that. I
added the if statement in case one of the places that do set it does
call this.
In other words, I added the if statement to make it more robust and
prevent a bug in the future ;-)
-- Steve
>
> Thank you,
>
> > ---
> > kernel/trace/trace.c | 33 ++++++++++++++++++++++-----------
> > 1 file changed, 22 insertions(+), 11 deletions(-)
> >
> > diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
> > index 1c370ffbe062..8775930aa545 100644
> > --- a/kernel/trace/trace.c
> > +++ b/kernel/trace/trace.c
> > @@ -4815,6 +4815,25 @@ static const struct seq_operations tracer_seq_ops = {
> > .show = s_show,
> > };
> >
> > +/*
> > + * Note, as iter itself can be allocated and freed in different
> > + * ways, this function is only used to free its content, and not
> > + * the iterator itself. The only requirement to all the allocations
> > + * is that it must zero all fields (kzalloc), as freeing works with
> > + * ethier allocated content or NULL.
> > + */
> > +static void free_trace_iter_content(struct trace_iterator *iter)
> > +{
> > + /* The fmt is either NULL, allocated or points to static_fmt_buf */
> > + if (iter->fmt != static_fmt_buf)
> > + kfree(iter->fmt);
> > +
> > + kfree(iter->temp);
> > + kfree(iter->buffer_iter);
> > + mutex_destroy(&iter->mutex);
> > + free_cpumask_var(iter->started);
> > +}
> > +
> > static struct trace_iterator *
> > __tracing_open(struct inode *inode, struct file *file, bool snapshot)
> > {
> > @@ -4922,8 +4941,7 @@ __tracing_open(struct inode *inode, struct file *file, bool snapshot)
> >
> > fail:
> > mutex_unlock(&trace_types_lock);
> > - kfree(iter->temp);
> > - kfree(iter->buffer_iter);
> > + free_trace_iter_content(iter);
> > release:
> > seq_release_private(inode, file);
> > return ERR_PTR(-ENOMEM);
> > @@ -5002,11 +5020,7 @@ static int tracing_release(struct inode *inode, struct file *file)
> >
> > mutex_unlock(&trace_types_lock);
> >
> > - mutex_destroy(&iter->mutex);
> > - free_cpumask_var(iter->started);
> > - kfree(iter->fmt);
> > - kfree(iter->temp);
> > - kfree(iter->buffer_iter);
> > + free_trace_iter_content(iter);
> > seq_release_private(inode, file);
> >
> > return 0;
> > @@ -6763,10 +6777,7 @@ static int tracing_release_pipe(struct inode *inode, struct file *file)
> >
> > mutex_unlock(&trace_types_lock);
> >
> > - free_cpumask_var(iter->started);
> > - kfree(iter->fmt);
> > - kfree(iter->temp);
> > - mutex_destroy(&iter->mutex);
> > + free_trace_iter_content(iter);
> > kfree(iter);
> >
> > trace_array_put(tr);
> > --
> > 2.40.1
>
>
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2023-07-28 23:08 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2023-07-15 14:12 [PATCH v2 0/2] tracing: Clean up how iter is freed Steven Rostedt
2023-07-15 14:12 ` [PATCH v2 1/2] tracing: Remove unnecessary copying of tr->current_trace Steven Rostedt
2023-07-15 14:12 ` [PATCH v2 2/2] tracing: Add free_trace_iter_content() helper function Steven Rostedt
2023-07-26 13:42 ` Masami Hiramatsu
2023-07-28 23:08 ` Steven Rostedt
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).