linux-kernel - [PATCH 0/2] iommu: Fix the longstanding probe issues

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: <cover.1739486121.git.robin.murphy@arm.com>
Date: Thu, 13 Feb 2025 23:48:58 +0000
From: Robin Murphy <robin.murphy@....com>
To: Lorenzo Pieralisi <lpieralisi@...nel.org>,
	Hanjun Guo <guohanjun@...wei.com>,
	Sudeep Holla <sudeep.holla@....com>,
	"Rafael J. Wysocki" <rafael@...nel.org>,
	Len Brown <lenb@...nel.org>,
	Russell King <linux@...linux.org.uk>,
	Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
	Danilo Krummrich <dakr@...nel.org>,
	Stuart Yoder <stuyoder@...il.com>,
	Laurentiu Tudor <laurentiu.tudor@....com>,
	Nipun Gupta <nipun.gupta@....com>,
	Nikhil Agarwal <nikhil.agarwal@....com>,
	Joerg Roedel <joro@...tes.org>,
	Will Deacon <will@...nel.org>,
	Rob Herring <robh@...nel.org>,
	Saravana Kannan <saravanak@...gle.com>,
	Bjorn Helgaas <bhelgaas@...gle.com>
Cc: linux-acpi@...r.kernel.org,
	linux-arm-kernel@...ts.infradead.org,
	linux-kernel@...r.kernel.org,
	iommu@...ts.linux.dev,
	devicetree@...r.kernel.org,
	linux-pci@...r.kernel.org,
	Charan Teja Kalla <quic_charante@...cinc.com>
Subject: [PATCH 0/2] iommu: Fix the longstanding probe issues

Hi all,

Finally, it's that thing I've been mentioning on and off for the last 5
years or so! After far too much thinking about "clean" ways to refactor
everything towards an eventual goal of working properly, I finally gave
up and decided to see how bad it would be to just get it working, then
worry about cleaning up afterwards. Frankly I still can't quiet believe
how small the patch turned out to be...

Patch #1 is a fix for what I think is the only "expected" race in the
core API design - we knew we had that window where a default domain is
NULL, but the potential consequences weren't clear until Charan figured
it out the hard way. That should be good to go already, and it's not
strictly a dependency for patch #2, but it seemed worth posting them
together on the common theme.

Patch #2 is then the first big bite out of the problem I feel compelled
to fix because it's my name on d7b0558230e4 ("iommu/of: Prepare for
deferred IOMMU configuration") which, in hindsight, started it. I hope
this will be the only patch which has to touch absolutely everything all
at once. I've not got very far with the follow-up patches yet, but I do
foresee being able to tackle the driver core, firmware subsystems and
bus drivers more independently. I don't expect anyone to love it, but I
hope you can at least see where I'm going and give a bit of leeway for
now :)

I also do rather expect that WARN_ON to be hit, but the intent is that
I'd like to get the attention of any of_dma_configure() abusers so they
can help me figure out what the heck their code *should* be doing...

Thanks,
Robin.

Robin Murphy (2):
  iommu: Handle race with default domain setup
  iommu: Get DT/ACPI parsing into the proper probe path

 drivers/acpi/arm64/dma.c        |  5 ++++
 drivers/acpi/scan.c             | 10 +++----
 drivers/amba/bus.c              |  2 +-
 drivers/base/platform.c         |  2 +-
 drivers/bus/fsl-mc/fsl-mc-bus.c |  2 +-
 drivers/cdx/cdx.c               |  2 +-
 drivers/iommu/iommu.c           | 48 +++++++++++++++++++++++++--------
 drivers/iommu/of_iommu.c        | 10 ++++++-
 drivers/of/device.c             |  7 ++++-
 drivers/pci/pci-driver.c        |  2 +-
 10 files changed, 65 insertions(+), 25 deletions(-)

-- 
2.39.2.101.g768bb238c484.dirty