linux-kernel - pcie-uniphier: race condition in masking/unmasking interrupts

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-next>] [day] [month] [year] [list]

Message-ID: <20210818113820.fzjeouy6tohbzuad@pali>
Date:   Wed, 18 Aug 2021 13:38:20 +0200
From:   Pali Rohár <pali@...nel.org>
To:     Kunihiko Hayashi <hayashi.kunihiko@...ionext.com>
Cc:     Marc Zyngier <maz@...nel.org>, linux-pci@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: pcie-uniphier: race condition in masking/unmasking interrupts

Hello!

Marc pointed during review of pci-aardvark patches one issue which I see
that is available also in the current pcie-uniphier.c driver.

When masking or unmasking interrupts there is read-modify-write sequence
for PCL_RCV_INTX_MASK_SHIFT register without any locking:

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/pci/controller/dwc/pcie-uniphier.c?h=v5.13#n171

So when trying to mask/unmask two interrupts at the same time there is
race condition as updating that PCL_RCV_INTX_MASK_SHIFT register is not
atomic.

Could you look at it?