Date: Fri, 15 Dec 2017 10:53:05 +0100
From: Peter Zijlstra <peterz@infradead.org>
To: Teng Qin <qinteng@fb.com>
Cc: Alexei Starovoitov <ast@fb.com>, "mingo@redhat.com" <mingo@redhat.com>,
        "bgregg@netflix.com" <bgregg@netflix.com>,
        "daniel@iogearbox.net" <daniel@iogearbox.net>,
        Yonghong Song <yhs@fb.com>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        Kernel Team <Kernel-team@fb.com>
Subject: Re: [PATCH tip 0/3] Improvements of scheduler related Tracepoints
Message-ID: <20171215095305.o7yinfsoym6bgaem@hirez.programming.kicks-ass.net>
References: <20171214202044.1629279-1-qinteng@fb.com>
 <20171214204932.GH3326@worktop>
 <1632e487-ee65-b50d-85e5-82f42c69fea1@fb.com>
 <20171215073908.myx3wgka7qimcmsg@hirez.programming.kicks-ass.net>
 <935FB65D-395D-4122-8E40-75CA04FF2111@fb.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <935FB65D-395D-4122-8E40-75CA04FF2111@fb.com>
User-Agent: NeoMutt/20170609 (1.8.3)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 701
Lines: 12

On Fri, Dec 15, 2017 at 08:53:30AM +0000, Teng Qin wrote:
> To have access to related task_struct is one of the main purposes of these
> patches. Take sched_switch as an example. We depend on the implementation
> of the Tracepoint is called from prev or next (which could, although unlikedly,
> change) and use current to get that task_struct, which feels, correct
> me if I'm wrong, kind of defeating the purpose of Tracepoints being more
> implementation-independent than kprobes. Then we have to figure out another
> Tracepoint or most likely a kprobe function to get the other (prev or next)
> task_struct.

Go read how tracepoints work. The tracepoint_probe things get you
exactly what you want.