Re: [PATCH 1/2] hung_task: Show the blocker task if the task is hung on mutex

[Waiman Long]

On 2/19/25 9:08 PM, Masami Hiramatsu (Google) wrote:
> On Wed, 19 Feb 2025 18:58:10 -0500
> Steven Rostedt <rostedt@xxxxxxxxxxx> wrote:
>
> > On Thu, 20 Feb 2025 08:09:08 +0900
> > Masami Hiramatsu (Google) <mhiramat@xxxxxxxxxx> wrote:
> >
> > > So something like this?
> > >
> > > unsigned int	block_flags;
> > > union {
> > > 	struct mutex	*mutex;
> > > 	struct rwsem	+rwsem;
> > > 	struct rtmutex	*rtmutex;
> > > } blocked_on;
> > >
> > > enum {
> > > 	BLOCKED_ON_MUTEX;
> > > 	BLOCKED_ON_RWSEM;
> > > 	BLOCKED_ON_RTMUTEX;
> > > 	BLOCKED_ON_IO;
> > > } block_reason;
> > >
> > > For the safety, we may anyway lock the task anyway, but that is the
> > > same as stacktrace.
> > Why not make it into a single entity?
> >
> > struct blocked_on {
> > 	unsigned int	flags;
> > 	union {
> > 		struct mutex	*mutex;
> > 		struct rwsem	*rwsem;
> > 		struct rtmutex	*rtmutex;
> > 	} blocked_on;
> > };
> Yes, and we also merge current mutex_waiter too.

I don't think we should merge mutex_waiter as it contains additional 
fields that are not useful to others. We don't want enlarge the size of 
task_struct unnecessarily.

Cheers,
Longman