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:   Thu, 1 Oct 2020 16:23:08 +0200
From:   Peter Zijlstra <peterz@infradead.org>
To:     Dave Hansen <dave.hansen@intel.com>
Cc:     "Liang, Kan" <kan.liang@linux.intel.com>,
        Stephane Eranian <eranian@google.com>,
        Ingo Molnar <mingo@redhat.com>,
        Arnaldo Carvalho de Melo <acme@kernel.org>,
        LKML <linux-kernel@vger.kernel.org>,
        Mark Rutland <mark.rutland@arm.com>,
        Alexander Shishkin <alexander.shishkin@linux.intel.com>,
        Jiri Olsa <jolsa@redhat.com>, Andi Kleen <ak@linux.intel.com>,
        kirill.shutemov@linux.intel.com,
        Michael Ellerman <mpe@ellerman.id.au>,
        benh@kernel.crashing.org, Paul Mackerras <paulus@samba.org>,
        Andy Lutomirski <luto@amacapital.net>
Subject: Re: [PATCH V8 1/4] perf/core: Add PERF_SAMPLE_DATA_PAGE_SIZE
Message-ID: <20201001142308.GR2628@hirez.programming.kicks-ass.net>
References: <20200921152653.3924-1-kan.liang@linux.intel.com>
 <20200921152653.3924-2-kan.liang@linux.intel.com>
 <CABPqkBRYzXH-76BZ3DdxYp7bdyPcr3_WxuxOsJw=1YPE9EwZaw@mail.gmail.com>
 <4e974520-6d0f-68af-7eb8-fa52d95ba77b@linux.intel.com>
 <35e875ba-2c04-8452-5105-ccacf72840d8@intel.com>
 <20200930173042.GD2628@hirez.programming.kicks-ass.net>
 <641fa138-23b7-8fdd-27f9-87e2fdd9fa91@intel.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <641fa138-23b7-8fdd-27f9-87e2fdd9fa91@intel.com>
Precedence: bulk

On Thu, Oct 01, 2020 at 06:49:36AM -0700, Dave Hansen wrote:
> On 9/30/20 10:30 AM, Peter Zijlstra wrote:
> > In general though; I think using ->active_mm is a mistake though. That
> > code should be doing something like:
> > 
> > 
> > 	mm = current->mm;
> > 	if (!mm)
> > 		mm = &init_mm;
> > 
> 
> I was hoping that using ->active_mm would give us the *actual* copy of
> the page tables that's loaded in CR3 for kernel thraeds.  But, there are
> few if any practical advantages of doing that at the moment.

Some of us hate active_mm with a passion and want to remove it entirely
(/me waves at amluto).

Also, if !current->mm, it had better not be accessing anything outside
of the kernel map, so &init_mm should cover that just fine. And given
the kernel maps are shared between all CR3s, they'll see the exact same
pagetables.