2022-05-17 15:34:26

by Yuanjun Gong

Subject: [PATCH 1/1] tracing: fix possible null pointer dereference

From: Gong Yuanjun <[email protected]>

In hist_trigger_elt_data_alloc(), elt_data is freed by
hist_elt_data_free() if kcalloc fails.

static int hist_trigger_elt_data_alloc(struct tracing_map_elt *elt)
{
...
elt_data->field_var_str = kcalloc(n_str, sizeof(char *), GFP_KERNEL);
if (!elt_data->field_var_str) {
hist_elt_data_free(elt_data);
return -EINVAL;
}
...}

In hist_elt_data_free() the elt_data->field_var_str field should be
checked before dereference.

Signed-off-by: Gong Yuanjun <[email protected]>
---
kernel/trace/trace_events_hist.c | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/kernel/trace/trace_events_hist.c b/kernel/trace/trace_events_hist.c
index 44db5ba9cabb..73177c9f94b2 100644
--- a/kernel/trace/trace_events_hist.c
+++ b/kernel/trace/trace_events_hist.c
@@ -1576,11 +1576,11 @@ static void hist_elt_data_free(struct hist_elt_data *elt_data)
{
unsigned int i;

- for (i = 0; i < elt_data->n_field_var_str; i++)
- kfree(elt_data->field_var_str[i]);
-
- kfree(elt_data->field_var_str);
-
+ if (elt_data->field_var_str) {
+ for (i = 0; i < elt_data->n_field_var_str; i++)
+ kfree(elt_data->field_var_str[i]);
+ kfree(elt_data->field_var_str);
+ }
kfree(elt_data->comm);
kfree(elt_data);
}
--
2.17.1

2022-05-18 19:29:38

by Steven Rostedt

Subject: Re: [PATCH 1/1] tracing: fix possible null pointer dereference

On Tue, 17 May 2022 17:57:23 +0800
Yuanjun Gong <[email protected]> wrote:

> From: Gong Yuanjun <[email protected]>
>
> In hist_trigger_elt_data_alloc(), elt_data is freed by
> hist_elt_data_free() if kcalloc fails.
>
> static int hist_trigger_elt_data_alloc(struct tracing_map_elt *elt)
> {
> ...
> elt_data->field_var_str = kcalloc(n_str, sizeof(char *), GFP_KERNEL);
> if (!elt_data->field_var_str) {

But at this position elt_data->n_field_var_str is zero.

> hist_elt_data_free(elt_data);
> return -EINVAL;
> }
> ...}
>
> In hist_elt_data_free() the elt_data->field_var_str field should be
> checked before dereference.
>
> Signed-off-by: Gong Yuanjun <[email protected]>
> ---
> kernel/trace/trace_events_hist.c | 10 +++++-----
> 1 file changed, 5 insertions(+), 5 deletions(-)
>
> diff --git a/kernel/trace/trace_events_hist.c b/kernel/trace/trace_events_hist.c
> index 44db5ba9cabb..73177c9f94b2 100644
> --- a/kernel/trace/trace_events_hist.c
> +++ b/kernel/trace/trace_events_hist.c
> @@ -1576,11 +1576,11 @@ static void hist_elt_data_free(struct hist_elt_data *elt_data)
> {
> unsigned int i;
>
> - for (i = 0; i < elt_data->n_field_var_str; i++)

This loop will not execute because n_field_var_str is zero, and 0 < 0 is
false.

> - kfree(elt_data->field_var_str[i]);
> -
> - kfree(elt_data->field_var_str);

freeing a NULL pointer is OK to do.

There is nothing wrong with the current code. NACK.

-- Steve

> -
> + if (elt_data->field_var_str) {
> + for (i = 0; i < elt_data->n_field_var_str; i++)
> + kfree(elt_data->field_var_str[i]);
> + kfree(elt_data->field_var_str);
> + }
> kfree(elt_data->comm);
> kfree(elt_data);
> }