Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <066d686946951e270e8fca127d8332c80b6cfac8.camel@gmail.com>
Subject: Re: [PATCH v4 0/6] iio: Add buffer write() support
From:   Nuno =?ISO-8859-1?Q?S=E1?= <noname.nuno@gmail.com>
To:     Jonathan Cameron <Jonathan.Cameron@Huawei.com>,
        Paul Cercueil <paul@crapouillou.net>
Cc:     Jonathan Cameron <jic23@kernel.org>,
        Lars-Peter Clausen <lars@metafoo.de>,
        Michael Hennerich <Michael.Hennerich@analog.com>,
        linux-iio@vger.kernel.org, linux-kernel@vger.kernel.org
Date:   Thu, 31 Aug 2023 13:01:24 +0200
In-Reply-To: <20230830171836.000045c3@Huawei.com>
References: <20230807112113.47157-1-paul@crapouillou.net>
         <20230830171118.00007726@Huawei.com> <20230830171836.000045c3@Huawei.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
User-Agent: Evolution 3.48.4 (3.48.4-1.fc38) 
MIME-Version: 1.0
Precedence: bulk

On Wed, 2023-08-30 at 17:18 +0100, Jonathan Cameron wrote:
> On Wed, 30 Aug 2023 17:11:18 +0100
> Jonathan Cameron <Jonathan.Cameron@Huawei.com> wrote:
>=20
> > On Mon,=C2=A0 7 Aug 2023 13:21:07 +0200
> > Paul Cercueil <paul@crapouillou.net> wrote:
> >=20
> > > [V3 was: "iio: new DMABUF based API, v3"][1]
> > >=20
> > > Hi Jonathan,
> > >=20
> > > This is a subset of my patchset that introduced a new interface based=
 on
> > > DMABUF objects [1]. It adds write() support to the IIO buffer
> > > infrastructure.
> > >=20
> > > The reason it is not the full IIO-DMABUF patchset, is because you
> > > requested performance benchmarks - and our current numbers are barely
> > > better (~ +10%) than the fileio interface. There is a good reason for
> > > that: V3 of the patchset switched from having the IIO core creating t=
he
> > > DMABUFs backed by physically contiguous memory, to having the IIO cor=
e
> > > being a simple DMABUF importer, and having the DMABUFs created
> > > externally. We now use the udmabuf driver to create those, and they a=
re
> > > allocated from paged memory. While this works perfectly fine, our
> > > buffers are now cut in 4 KiB chunks (pages), non-contiguous in memory=
,
> > > which causes the DMA hardware to create an IRQ storm, as it raises an
> > > interrupt after each 4 KiB in the worst case scenario.=C2=A0=20
> >=20
> > Interesting. I'm guessing you don't necessarily need contiguous memory
> > and huge pages would get rid of most of that overhead?
> >=20
> > Given embedded target those huge pages are hard to get so you need
> > hugetlb support to improve the chances of it working.=C2=A0 Some quick =
searching
> > suggests there is possible support on the way.
> > https://lore.kernel.org/linux-mm/20230817064623.3424348-1-vivek.kasired=
dy@intel.com/
> >=20
> >=20
> > >=20
> > > Anyway, this is not directly a problem of the IIO-DMABUF code - but I
> > > can't really upstream a shiny new interface that I claim is much fast=
er,
> > > without giving numbers.
> > >=20
> > > So while we fix this (either by updating the DMA IP and driver to
> > > support scatter-gather)=C2=A0=20
> >=20
> > Long run you almost always end up needing that unless contig requiremen=
ts
> > are small and you want a robust solution.=C2=A0 I'm guessing no IOMMU t=
o pretend
> > it's all contiguous...=20
> >=20
> > > or by hacking something quick to give us
> > > physically contiguous DMABUFs just for the benchmark), I thought it
> > > would make sense to upstream the few patches of the V3 patchset that =
are
> > > needed for the IIO-DMABUF interface but aren't directly related.=C2=
=A0=20
> >=20
> > Good idea.
> >=20
> > >=20
> > > As for write() support, Nuno (Cc'd) said he will work on upstreaming =
the
> > > DAC counterpart of adc/adi-axi-adc.c in the next few weeks, so there
> > > will be a user for the buffer write() support. I hope you are okay wi=
th
> > > this - otherwise, we can just wait until this work is done and submit=
 it
> > > all at once.=C2=A0=20
> >=20
> > Absolutely fine, though I won't pick this up without the user also bein=
g
> > ready to go.
>=20
>=20
> Having looked through these again, they are straight forward so no change=
s
> requested from me.=C2=A0 Nuno, if you can add this set into appropriate
> point in your series that will make use of it that will make my life easi=
er
> and ensure and minor rebasing etc happens without having to bother Paul.
>=20

Sure...

- Nuno S=C3=A1
>=20