linux-kernel - Re: PCI: Race condition in 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] [thread-next>] [day] [month] [year] [list]

Message-ID: <20220120230050.GA1077958@bhelgaas>
Date:   Thu, 20 Jan 2022 17:00:50 -0600
From:   Bjorn Helgaas <helgaas@...nel.org>
To:     Krzysztof Hałasa <khalasa@...p.pl>
Cc:     Pali Rohár <pali@...nel.org>,
        Bjorn Helgaas <bhelgaas@...gle.com>, linux-pci@...r.kernel.org,
        linux-kernel@...r.kernel.org,
        Lorenzo Pieralisi <lorenzo.pieralisi@....com>,
        Gregory Clement <gregory.clement@...tlin.com>,
        Andrew Lunn <andrew@...n.ch>,
        Krzysztof Wilczyński <kw@...ux.com>
Subject: Re: PCI: Race condition in pci_create_sysfs_dev_files (can't boot)

On Fri, Jul 30, 2021 at 10:18:44AM +0200, Krzysztof Hałasa wrote:
> Pali, et al,
> 
> I'm encountering a problem booting an i.MX6-based device (Gateworks
> Ventana SBC). This is apparently a known issue:
> https://lkml.org/lkml/2020/7/16/388
> 
> Do you guys know of a fix for this? Booting this machine reliably is
> lately impossible.

Hi Krzysztof, is this still an issue?

e1d3f3268b0e ("PCI/sysfs: Convert "config" to static attribute") and
similar patches appeared in v5.13, and they should resolve things like
this:

  sysfs: cannot create duplicate filename '/devices/platform/soc/d0070000.pcie/pci0000:00/0000:00:00.0/config'

But your report is different because it looks like we have a problem
when creating a *directory*:

> sysfs: cannot create duplicate filename '/devices/platform/soc/1ffc000.pcie/pci0000:00/0000:00:00.0/0000:e0'

and then it seems that we corrupted a pointer:

> Unable to handle kernel paging request at virtual address 6f736572 (ASCII = "reso")

which is even more worrisome.

Bjorn