Re: [PATCH v2] perf top: expand the range of multithreaded phase

[Namhyung Kim]

Hello,

On Wed, Apr 5, 2023 at 7:54 PM Hangliang Lai <laihangliang1@xxxxxxxxxx> wrote:
>
> In __cmd_top, perf_set_multithreaded is used to enable pthread_rwlock, thus
> donw_read and down_write can work to handle concurrency problems. Then top
> use perf_set_singlethreaded and switch to single threaded phase, assuming
> that no thread concurrency will happen later.
> However, a UAF problem could occur in perf top in single threaded phase,
> The concurrent procedure is like this:
> display_thread                              process_thread
> --------------                              --------------
> thread__comm_len
>   -> thread__comm_str
>     -> __thread__comm_str(thread)
>                                             thread__delete
>                                              -> comm__free
>                                               -> comm_str__put
>                                                -> zfree(&cs->str)
>     -> thread->comm_len = strlen(comm);
> Since in single thread phase, perf_singlethreaded is true, down_read and
> down_write can not work to avoid concurrency problems.
> This patch put perf_set_singlethreaded to the function tail to expand the
> multithreaded phase range, make display_thread and process_thread run
> safe.
>
> Signed-off-by: Hangliang Lai  <laihangliang1@xxxxxxxxxx>
> Reviewed-by: Yunfeng Ye <yeyunfeng@xxxxxxxxxx>
> ---
> v1 -> v2
>  - Since perf top is always multi-threaded, remove top->nr_threads_synthesize judgment.

Not always, the synthesis can run in a single thread.

>
>  tools/perf/builtin-top.c | 9 ++++-----
>  1 file changed, 4 insertions(+), 5 deletions(-)
>
> diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c
> index d4b5b02bab73..a18db1ee87fa 100644
> --- a/tools/perf/builtin-top.c
> +++ b/tools/perf/builtin-top.c
> @@ -1242,8 +1242,7 @@ static int __cmd_top(struct perf_top *top)
>         if (perf_session__register_idle_thread(top->session) < 0)
>                 return ret;
>
> -       if (top->nr_threads_synthesize > 1)
> -               perf_set_multithreaded();
> +       perf_set_multithreaded();

I think this part should be kept as is.

>
>         init_process_thread(top);
>
> @@ -1273,9 +1272,6 @@ static int __cmd_top(struct perf_top *top)
>                                     top->evlist->core.threads, true, false,
>                                     top->nr_threads_synthesize);
>
> -       if (top->nr_threads_synthesize > 1)
> -               perf_set_singlethreaded();

Here you can make it multi-threaded unconditionally.

Thanks,
Namhyung

> -
>         if (perf_hpp_list.socket) {
>                 ret = perf_env__read_cpu_topology_map(&perf_env);
>                 if (ret < 0) {
> @@ -1352,6 +1348,9 @@ static int __cmd_top(struct perf_top *top)
>  out_join_thread:
>         cond_signal(&top->qe.cond);
>         pthread_join(thread_process, NULL);
> +
> +       perf_set_singlethreaded();
> +
>         return ret;
>  }
>
> --
> 2.33.0
>