From:   Alexander Stein <alexander.stein@ew.tq-group.com>
To:     Bjorn Helgaas <bhelgaas@google.com>,
        Oliver Neukum <oneukum@suse.com>,
        Oliver Neukum <oneukum@suse.com>
Cc:     Korneliusz Osmenda <korneliuszo@gmail.com>,
        linux-pci@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2 1/1] Guard pci_create_sysfs_dev_files with atomic value
Date:   Thu, 16 Mar 2023 16:00:42 +0100
Message-ID: <3554992.taCxCBeP46@steina-w>
Organization: TQ-Systems GmbH
In-Reply-To: <ced8713f-69de-e48a-37eb-4f844e651b6b@suse.com>
References: <20230316091540.494366-1-alexander.stein@ew.tq-group.com> <3607385.usQuhbGJ8B@steina-w> <ced8713f-69de-e48a-37eb-4f844e651b6b@suse.com>
MIME-Version: 1.0
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain; charset="iso-8859-1"
Precedence: bulk

Hi Oliver,

Am Donnerstag, 16. M=E4rz 2023, 15:01:25 CET schrieb Oliver Neukum:
> On 16.03.23 14:16, Alexander Stein wrote:
> > But isn't the root bridge discovered by the driver (pci-imx6 in this ca=
se)
> > for that? And the driver probe path eventually calls into the sysfs file
> > creation. I compared the file creation to usb, as this is a discoverable
> > bus as well. There is no special initialization regarding sysfs.
>=20
> If you discover a bus system you always have the option of creating of
> virtual hotplug event for the root hub or host controller.
> But for PCI that is a bad design choice. USB is different.

I'm not sure if I can follow you here. Can you elaborate?

> > If, for some reason, the device enumeration for PCI bus during
> > imx6_pcie_probe is delayed after pci_sysfs_init initcall, this initcall
> > essentially does nothing, no devices or busses to iterate. Which means
> > the complete pcie sysfs
> On your specific system. You cannot use that as a model for all systems.

I am aware that my platform is not a role model for the others. But I've ye=
t=20
to get information what is actually different on other platforms.

> > creation is done from bridge probe path. There is no reason to iterate
> > over
> > discovered PCIe devices/busses separately.
>=20
> If there is no other PCI device, the loop is a nop. But otherwise it is
> necessary.

How is it necessary? How do these PCI devices get attaches to the pci_bus_t=
ype=20
bus without calling pci_bus_add_device?

> >>> So technically the device is not probed from within a initcall but a
> >>> kthread. It is set to be probed asynchronous in imx6_pcie_driver.
> >>=20
> >> That may be the problem, respectively that system is incomplete
> >> You are registering a PCI bridge. The PCI subsystem should be
> >> done setting up when you run. That is just a simple dependency.
> >=20
> > Is there such an dependency in the first place? I can't see anything, e=
ven
> > the late_initcall to pci_resource_alignment_sysfs_init is a different
> > matter.
> On your hardware, yes. In the kernel, no.
> That is the very point. The kernel is missing a way to represent a
> dependency.

Okay, so which dependency is provided by pci_sysfs_init, which are required=
 by=20
drivers then?

Best regards,
Alexander
=2D-=20
TQ-Systems GmbH | M=FChlstra=DFe 2, Gut Delling | 82229 Seefeld, Germany
Amtsgericht M=FCnchen, HRB 105018
Gesch=E4ftsf=FChrer: Detlef Schneider, R=FCdiger Stahl, Stefan Schneider
http://www.tq-group.com/