Before current_user_event_group(), it has allocated memory and save it
in @name, this should freed before return error.
Fixes: e5d271812e7a ("tracing/user_events: Move pages/locks into groups to prepare for namespaces")
Signed-off-by: Xiu Jianfeng <[email protected]>
---
kernel/trace/trace_events_user.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/kernel/trace/trace_events_user.c b/kernel/trace/trace_events_user.c
index ae78c2d53c8a..539b08ae7020 100644
--- a/kernel/trace/trace_events_user.c
+++ b/kernel/trace/trace_events_user.c
@@ -1100,8 +1100,10 @@ static int user_event_create(const char *raw_command)
group = current_user_event_group();
- if (!group)
+ if (!group) {
+ kfree(name);
return -ENOENT;
+ }
mutex_lock(&group->reg_mutex);
--
2.17.1
On Tue, 15 Nov 2022 09:44:45 +0800
Xiu Jianfeng <[email protected]> wrote:
> Before current_user_event_group(), it has allocated memory and save it
> in @name, this should freed before return error.
>
> Fixes: e5d271812e7a ("tracing/user_events: Move pages/locks into groups to prepare for namespaces")
> Signed-off-by: Xiu Jianfeng <[email protected]>
This looks good to me.
Acked-by: Masami Hiramatsu (Google) <[email protected]>
Thanks,
> ---
> kernel/trace/trace_events_user.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/kernel/trace/trace_events_user.c b/kernel/trace/trace_events_user.c
> index ae78c2d53c8a..539b08ae7020 100644
> --- a/kernel/trace/trace_events_user.c
> +++ b/kernel/trace/trace_events_user.c
> @@ -1100,8 +1100,10 @@ static int user_event_create(const char *raw_command)
>
> group = current_user_event_group();
>
> - if (!group)
> + if (!group) {
> + kfree(name);
> return -ENOENT;
> + }
>
> mutex_lock(&group->reg_mutex);
>
> --
> 2.17.1
>
--
Masami Hiramatsu (Google) <[email protected]>
On Wed, Nov 16, 2022 at 12:07:17AM +0900, Masami Hiramatsu wrote:
> On Tue, 15 Nov 2022 09:44:45 +0800
> Xiu Jianfeng <[email protected]> wrote:
>
> > Before current_user_event_group(), it has allocated memory and save it
> > in @name, this should freed before return error.
> >
Nice catch!
Acked-by: Beau Belgrave <[email protected]>
Thanks,
-Beau
> > Fixes: e5d271812e7a ("tracing/user_events: Move pages/locks into groups to prepare for namespaces")
> > Signed-off-by: Xiu Jianfeng <[email protected]>
>
> This looks good to me.
>
> Acked-by: Masami Hiramatsu (Google) <[email protected]>
>
> Thanks,
>
> > ---
> > kernel/trace/trace_events_user.c | 4 +++-
> > 1 file changed, 3 insertions(+), 1 deletion(-)
> >
> > diff --git a/kernel/trace/trace_events_user.c b/kernel/trace/trace_events_user.c
> > index ae78c2d53c8a..539b08ae7020 100644
> > --- a/kernel/trace/trace_events_user.c
> > +++ b/kernel/trace/trace_events_user.c
> > @@ -1100,8 +1100,10 @@ static int user_event_create(const char *raw_command)
> >
> > group = current_user_event_group();
> >
> > - if (!group)
> > + if (!group) {
> > + kfree(name);
> > return -ENOENT;
> > + }
> >
> > mutex_lock(&group->reg_mutex);
> >
> > --
> > 2.17.1
> >
>
>
> --
> Masami Hiramatsu (Google) <[email protected]>