Received-SPF: pass (google.com: domain of linux-kernel+bounces-91780-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:40f1:3f00::1 as permitted sender) client-ip=2604:1380:40f1:3f00::1;
Message-ID: <ca0e141e624edb4f9977ea9f805f74645654bca7.camel@gmail.com>
Subject: Re: [PATCH v3 0/2] Synchronize DT overlay removal with devlink
 removals
From: Nuno =?ISO-8859-1?Q?S=E1?= <noname.nuno@gmail.com>
To: Saravana Kannan <saravanak@google.com>, Rob Herring <robh@kernel.org>
Cc: Herve Codina <herve.codina@bootlin.com>, Greg Kroah-Hartman
 <gregkh@linuxfoundation.org>, "Rafael J. Wysocki" <rafael@kernel.org>,
 Frank Rowand <frowand.list@gmail.com>, Lizhi Hou <lizhi.hou@amd.com>, Max
 Zhen <max.zhen@amd.com>,  Sonal Santan <sonal.santan@amd.com>, Stefano
 Stabellini <stefano.stabellini@xilinx.com>, Jonathan Cameron
 <Jonathan.Cameron@huawei.com>, linux-kernel@vger.kernel.org, 
 devicetree@vger.kernel.org, Allan Nielsen <allan.nielsen@microchip.com>, 
 Horatiu Vultur <horatiu.vultur@microchip.com>, Steen Hegelund
 <steen.hegelund@microchip.com>, Luca Ceresoli <luca.ceresoli@bootlin.com>,
 Nuno Sa <nuno.sa@analog.com>, Thomas Petazzoni
 <thomas.petazzoni@bootlin.com>
Date: Tue, 05 Mar 2024 08:17:38 +0100
In-Reply-To: <CAGETcx96Nj1RnR2wFOZYJRr=kyBLkcobY0crxBFSLGWmNaXT4A@mail.gmail.com>
References: <20240229105204.720717-1-herve.codina@bootlin.com>
	 <20240304150248.GA211460-robh@kernel.org>
	 <CAGETcx96Nj1RnR2wFOZYJRr=kyBLkcobY0crxBFSLGWmNaXT4A@mail.gmail.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
User-Agent: Evolution 3.50.4 
Precedence: bulk
MIME-Version: 1.0

On Mon, 2024-03-04 at 22:17 -0800, Saravana Kannan wrote:
> On Mon, Mar 4, 2024 at 7:02=E2=80=AFAM Rob Herring <robh@kernel.org> wrot=
e:
> >=20
> > On Thu, Feb 29, 2024 at 11:52:01AM +0100, Herve Codina wrote:
> > > Hi,
> >=20
> > Please CC Saravana on this.
>=20
> Nuno, this is why I was replying to the older series. I didn't even
> get this one.

Arghh, I see... In lot's of replies I was mentioning you :)

- Nuno S=C3=A1

>=20
> >=20
> > >=20
> > > In the following sequence:
> > > =C2=A0 of_platform_depopulate(); /* Remove devices from a DT overlay =
node */
> > > =C2=A0 of_overlay_remove(); /* Remove the DT overlay node itself */
> > >=20
> > > Some warnings are raised by __of_changeset_entry_destroy() which=C2=
=A0 was
> > > called from of_overlay_remove():
> > > =C2=A0 ERROR: memory leak, expected refcount 1 instead of 2 ...
> > >=20
> > > The issue is that, during the device devlink removals triggered from =
the
> > > of_platform_depopulate(), jobs are put in a workqueue.
> > > These jobs drop the reference to the devices. When a device is no mor=
e
> > > referenced (refcount =3D=3D 0), it is released and the reference to i=
ts
> > > of_node is dropped by a call to of_node_put().
> > > These operations are fully correct except that, because of the
> > > workqueue, they are done asynchronously with respect to function call=
s.
> > >=20
> > > In the sequence provided, the jobs are run too late, after the call t=
o
> > > __of_changeset_entry_destroy() and so a missing of_node_put() call is
> > > detected by __of_changeset_entry_destroy().
> > >=20
> > > This series fixes this issue introducing device_link_wait_removal() i=
n
> > > order to wait for the end of jobs execution (patch 1) and using this
> > > function to synchronize the overlay removal with the end of jobs
> > > execution (patch 2).
> > >=20
> > > Compared to the previous iteration:
> > > =C2=A0
> > > https://lore.kernel.org/linux-kernel/20231130174126.688486-1-herve.co=
dina@bootlin.com/
> > > this v3 series:
> > > - add the missing device.h
> > >=20
> > > This series handles cases reported by Luca [1] and Nuno [2].
> > > =C2=A0 [1]: https://lore.kernel.org/all/20231220181627.341e8789@booty=
/
> > > =C2=A0 [2]:
> > > https://lore.kernel.org/all/20240205-fix-device-links-overlays-v2-2-5=
344f8c79d57@analog.com/
> > >=20
> > > Best regards,
> > > Herv=C3=A9
> > >=20
> > > Changes v2 -> v3
> > > =C2=A0 - Patch 1
> > > =C2=A0=C2=A0=C2=A0 No changes
> > >=20
> > > =C2=A0 - Patch 2
> > > =C2=A0=C2=A0=C2=A0 Add missing device.h
> > >=20
> > > Changes v1 -> v2
> > > =C2=A0 - Patch 1
> > > =C2=A0=C2=A0=C2=A0 Rename the workqueue to 'device_link_wq'
> > > =C2=A0=C2=A0=C2=A0 Add 'Fixes' tag and Cc stable
> > >=20
> > > =C2=A0 - Patch 2
> > > =C2=A0=C2=A0=C2=A0 Add device.h inclusion.
> > > =C2=A0=C2=A0=C2=A0 Call device_link_wait_removal() later in the overl=
ay removal
> > > =C2=A0=C2=A0=C2=A0 sequence (i.e. in free_overlay_changeset() functio=
n).
> > > =C2=A0=C2=A0=C2=A0 Drop of_mutex lock while calling device_link_wait_=
removal().
> > > =C2=A0=C2=A0=C2=A0 Add=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 'Fixes' ta=
g and Cc stable
> > >=20
> > > Herve Codina (2):
> > > =C2=A0 driver core: Introduce device_link_wait_removal()
> > > =C2=A0 of: overlay: Synchronize of_overlay_remove() with the devlink =
removals
> > >=20
> > > =C2=A0drivers/base/core.c=C2=A0=C2=A0=C2=A0 | 26 ++++++++++++++++++++=
+++---
> > > =C2=A0drivers/of/overlay.c=C2=A0=C2=A0 | 10 +++++++++-
> > > =C2=A0include/linux/device.h |=C2=A0 1 +
> > > =C2=A03 files changed, 33 insertions(+), 4 deletions(-)
> > >=20
> > > --
> > > 2.43.0
> > >=20