Received-SPF: softfail (google.com: domain of transitioning linux-kernel-owner@vger.kernel.org does not designate 23.128.96.19 as permitted sender) client-ip=23.128.96.19;
Date:   Thu, 21 Apr 2022 13:35:55 +1000
From:   Nicholas Piggin <npiggin@gmail.com>
Subject: Re: [PATCH v2 bpf 1/3] vmalloc: replace VM_NO_HUGE_VMAP with
 VM_ALLOW_HUGE_VMAP
To:     Christoph Hellwig <hch@infradead.org>, Song Liu <song@kernel.org>
Cc:     Andrew Morton <akpm@linux-foundation.org>,
        Andrii Nakryiko <andrii@kernel.org>,
        Alexei Starovoitov <ast@kernel.org>, bpf <bpf@vger.kernel.org>,
        Daniel Borkmann <daniel@iogearbox.net>, imbrenda@linux.ibm.com,
        open list <linux-kernel@vger.kernel.org>,
        Linux-MM <linux-mm@kvack.org>,
        Luis Chamberlain <mcgrof@kernel.org>,
        rick.p.edgecombe@intel.com
References: <20220411233549.740157-1-song@kernel.org>
        <20220411233549.740157-2-song@kernel.org> <YlT9i9DFvwDx9+AD@infradead.org>
        <CAPhsuW7XJHa3OaTT-4=33c70gUjCaWFrVe8h8J-hZetjxXeeog@mail.gmail.com>
        <1650507506.z839xl6pvt.astroid@bobo.none>
In-Reply-To: <1650507506.z839xl6pvt.astroid@bobo.none>
MIME-Version: 1.0
Message-Id: <1650512125.tnay4e9v4h.astroid@bobo.none>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

Excerpts from Nicholas Piggin's message of April 21, 2022 12:24 pm:
> Excerpts from Song Liu's message of April 12, 2022 4:00 pm:
>> On Mon, Apr 11, 2022 at 9:18 PM Christoph Hellwig <hch@infradead.org> wr=
ote:
>>>
>>> On Mon, Apr 11, 2022 at 04:35:46PM -0700, Song Liu wrote:
>>> > Huge page backed vmalloc memory could benefit performance in many cas=
es.
>>> > Since some users of vmalloc may not be ready to handle huge pages,
>>> > VM_NO_HUGE_VMAP was introduced to allow vmalloc users to opt-out huge
>>> > pages. However, it is not easy to add VM_NO_HUGE_VMAP to all the user=
s
>>> > that may try to allocate >=3D PMD_SIZE pages, but are not ready to ha=
ndle
>>> > huge pages properly.
>>>
>>> This is a good place to document what the problems are, and how they ar=
e
>>> hard to track down (e.g. because the allocations are passed down I/O
>>> stacks)
>>=20
>> Will add it in v3.
>>=20
>>>
>>> >
>>> > Replace VM_NO_HUGE_VMAP with an opt-in flag, VM_ALLOW_HUGE_VMAP, so t=
hat
>>> > users that benefit from huge pages could ask specificially.
>>> >
>>> > Also, replace vmalloc_no_huge() with opt-in helper vmalloc_huge().
>>>
>>> We still need to find out what the primary users of the large vmalloc
>>> hashes was and convert them.
>>=20
>> @ Claudio and Nicholas,
>>=20
>> Could you please help identify users of large vmalloc? So far, I found
>> alloc_large_system_hash(), and something like the following seems to
>> work:
>=20
> The large system hashes were the main ones I was interested in. IIRC=20
> there was a few more in some drivers or tracing things depending on
> config but those are less important (to me at least).

Oh there is also a reverse map array in KVM now I think of it.

Thanks,
Nick