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;
MIME-Version: 1.0
References: <20210526080741.GW30378@techsingularity.net> <YK9SiLX1E1KAZORb@infradead.org>
 <20210527090422.GA30378@techsingularity.net> <YK9j3YeMTZ+0I8NA@infradead.org>
 <CAEf4BzZLy0s+t+Nj9QgUNM66Ma6HN=VkS+ocgT5h9UwanxHaZQ@mail.gmail.com>
 <CAEf4BzbzPK-3cyLFM8QKE5-o_dL7=UCcvRF+rEqyUcHhyY+FJg@mail.gmail.com> <8fe547e9e87f40aebce82021d76a2d08@AcuMS.aculab.com>
In-Reply-To: <8fe547e9e87f40aebce82021d76a2d08@AcuMS.aculab.com>
From:   Andrii Nakryiko <andrii.nakryiko@gmail.com>
Date:   Sat, 29 May 2021 17:46:54 -0700
Message-ID: <CAEf4Bza_yDe7=0jNuPcFnJOTiHPTWD43cHyQqqa10i228P_OVw@mail.gmail.com>
Subject: Re: [PATCH] mm/page_alloc: Work around a pahole limitation with
 zero-sized struct pagesets
To:     David Laight <David.Laight@aculab.com>
Cc:     Christoph Hellwig <hch@infradead.org>,
        Mel Gorman <mgorman@techsingularity.net>,
        Andrew Morton <akpm@linux-foundation.org>,
        Arnaldo Carvalho de Melo <acme@redhat.com>,
        Michal Suchanek <msuchanek@suse.de>,
        Alexei Starovoitov <ast@kernel.org>,
        Daniel Borkmann <daniel@iogearbox.net>,
        Martin KaFai Lau <kafai@fb.com>,
        Song Liu <songliubraving@fb.com>, Yonghong Song <yhs@fb.com>,
        John Fastabend <john.fastabend@gmail.com>,
        KP Singh <kpsingh@kernel.org>,
        open list <linux-kernel@vger.kernel.org>,
        Jiri Olsa <jolsa@kernel.org>,
        Hritik Vijay <hritikxx8@gmail.com>, bpf <bpf@vger.kernel.org>,
        Linux-Net <netdev@vger.kernel.org>, Linux-MM <linux-mm@kvack.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Fri, May 28, 2021 at 1:09 AM David Laight <David.Laight@aculab.com> wrote:
>
> From: Andrii Nakryiko
> > Sent: 27 May 2021 15:42
> ...
> > I agree that empty structs are useful, but here we are talking about
> > per-CPU variables only, which is the first use case so far, as far as
> > I can see. If we had pahole 1.22 released and widely packaged it could
> > have been a viable option to force it on everyone.
> ...
>
> Would it be feasible to put the sources for pahole into the
> kernel repository and build it at the same time as objtool?
>
> That would remove any issues about the latest version
> not being available.

That would be great for the kernel build, but pahole is more than just
a DWARF-to-BTF converter and it has a substantial amount of logic for
loading and processing DWARF before it gets converted to BTF. All
BTF-related pieces are provided by libbpf, which is already part of
kernel sources, so that's not a problem. DWARF processing is a problem
and would add a new dependency on libdw-devel, at least.

>
>         David
>
> -
> Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
> Registration No: 1397386 (Wales)