Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Wed, 3 Feb 2021 14:37:05 +0100
From:   Peter Zijlstra <peterz@infradead.org>
To:     Dmitry Vyukov <dvyukov@google.com>
Cc:     Ingo Molnar <mingo@redhat.com>,
        Arnaldo Carvalho de Melo <acme@kernel.org>,
        Mark Rutland <mark.rutland@arm.com>,
        Alexander Shishkin <alexander.shishkin@linux.intel.com>,
        Jiri Olsa <jolsa@redhat.com>,
        Namhyung Kim <namhyung@kernel.org>,
        Will Deacon <will@kernel.org>,
        LKML <linux-kernel@vger.kernel.org>,
        Matt Morehouse <mascasa@google.com>
Subject: Re: Process-wide watchpoints
Message-ID: <YBqnAYVdNM4uyGny@hirez.programming.kicks-ass.net>
References: <CACT4Y+YPrXGw+AtESxAgPyZ84TYkNZdP0xpocX2jwVAbZD=-XQ@mail.gmail.com>
 <20201112103125.GV2628@hirez.programming.kicks-ass.net>
 <CACT4Y+ayRHua-6UyRwSM3=_oi+NkXbaO3-zZ1mpDmWonbybkeA@mail.gmail.com>
 <CACT4Y+bW1gpv8bz0vswaVUt-OB07oJ3NBeTi+vchAe8TTWK+mg@mail.gmail.com>
 <CACT4Y+ZsKXfAxrzJGQc5mJ+QiP5sAw7zKWtciS+07qZzSf33mw@mail.gmail.com>
 <CACT4Y+YeRtOTsMQ8xxZg-=nbv+yuJvYYhBErT46M8jtSHmiw6g@mail.gmail.com>
 <YBqXPmbpXf4hnlj3@hirez.programming.kicks-ass.net>
 <CACT4Y+a-9kqX0ZkNz-ygib+ERn41HVo_8Wx6oYMQmPjTC06j7g@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <CACT4Y+a-9kqX0ZkNz-ygib+ERn41HVo_8Wx6oYMQmPjTC06j7g@mail.gmail.com>
Precedence: bulk

On Wed, Feb 03, 2021 at 01:49:56PM +0100, Dmitry Vyukov wrote:
> On Wed, Feb 3, 2021 at 1:29 PM Peter Zijlstra <peterz@infradead.org> wrote:
> >
> > On Mon, Feb 01, 2021 at 09:50:20AM +0100, Dmitry Vyukov wrote:
> > > Or, alternatively would it be reasonable for perf to generate SIGTRAP
> > > directly on watchpoint hit (like ptrace does)? That's what I am
> > > ultimately trying to do by attaching a bpf program.
> >
> > Perf should be able to generate signals, The perf_event_open manpage
> > lists two ways of trigering signals. The second way doesn't work for
> > you, due to it not working on inherited counters, but would the first
> > work?
> >
> > That is, set attr::wakeup_events and fcntl(F_SETSIG).
> 
> The problem is that this sends a signal to the fd owner rather than
> the thread that hit the breakpoint. At least that's what happened in
> our tests. We would like to send a signal to the thread that hit the
> breakpoint.

Ah indeed.. all of this was aimed at self-monitoring.

Letting perf send a signal to the monitored task is intrusive.. let me
think on that.