linux-kernel - Re: [PATCH v2 1/1] Guard pci_create_sysfs_dev

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Tue, 21 Mar 2023 10:09:22 +0100
From:   Oliver Neukum <oneukum@...e.com>
To:     Oliver Neukum <oneukum@...e.com>,
        Alexander Stein <alexander.stein@...tq-group.com>,
        Bjorn Helgaas <bhelgaas@...gle.com>
Cc:     Korneliusz Osmenda <korneliuszo@...il.com>,
        linux-pci@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 1/1] Guard pci_create_sysfs_dev_files with atomic value

Am Donnerstag, 16. März 2023, 15:01:25 CET schrieb Oliver Neukum:
> I'm not sure if I can follow you here. Can you elaborate?

There are far better reasons to leave the setup of a PCI bus as the firmware
has done it than to leave a USB as is.

> How is it necessary? How do these PCI devices get attaches to the pci_bus_type 
> bus without calling pci_bus_add_device?

AFAICT they don't. But somebody has to call it.

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

It isn't. It is missing from the code. But it exists in reality. That is the point.
You have a race condition between two probes. We cannot have that.
Hence IMHO the dependency would be best expressed by waiting for pci_sysfs_init()
to finish in the init sequence before you add any PCI bridges or devices to the system.

	Regards
		Oliver