From: Nicolin Chen <nicolinc@nvidia.com>
To: <will@kernel.org>, <robin.murphy@arm.com>, <joro@8bytes.org>,
<bhelgaas@google.com>, <jgg@nvidia.com>
Cc: <rafael@kernel.org>, <lenb@kernel.org>, <praan@google.com>,
<kees@kernel.org>, <baolu.lu@linux.intel.com>,
<smostafa@google.com>, <Alexander.Grest@microsoft.com>,
<kevin.tian@intel.com>, <miko.lenczewski@arm.com>,
<linux-arm-kernel@lists.infradead.org>, <iommu@lists.linux.dev>,
<linux-kernel@vger.kernel.org>, <linux-acpi@vger.kernel.org>,
<linux-pci@vger.kernel.org>, <vsethi@nvidia.com>
Subject: [PATCH v1 0/2] iommu/arm-smmu-v3: Reset PCI device upon ATC invalidate timeout
Date: Wed, 4 Mar 2026 21:21:40 -0800 [thread overview]
Message-ID: <cover.1772686998.git.nicolinc@nvidia.com> (raw)
This is a small series to handle ATC invalidate timeout in the CMDQ Error
handler. When the ATC is out of sync, the device might corrupt the memory
due to the stale cache entries. The safest way is to reset the device for
a recovery that wipes out the ATC as well.
On the other hand, a reset attempt in the ATC recovery thread might fail,
which leaves the IOMMU driver no choice but to isolate the device.
pci_dev_reset_iommu_prepare() was introduced previously to fence the ATS
and ATC during a PCI reset. It's easy to use yet doesn't allow a re-entry
nor support the recovery very well.
In this series,
- loosen the re-entry and block ATS if reset fails
- add a reset routine in SMMUv3 driver
This is on Github:
https://github.com/nicolinc/iommufd/commits/smmuv3_atc_timeout
Nicolin Chen (2):
iommu: Do not call pci_dev_reset_iommu_done() unless reset succeeds
iommu/arm-smmu-v3: Recover ATC invalidate timeouts
drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.h | 5 +
drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c | 131 +++++++++++++++++++-
drivers/iommu/iommu.c | 16 ++-
drivers/pci/pci-acpi.c | 11 +-
drivers/pci/pci.c | 50 +++++++-
drivers/pci/quirks.c | 11 +-
6 files changed, 207 insertions(+), 17 deletions(-)
--
2.43.0
next reply other threads:[~2026-03-05 5:22 UTC|newest]
Thread overview: 44+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-03-05 5:21 Nicolin Chen [this message]
2026-03-05 5:21 ` [PATCH v1 1/2] iommu: Do not call pci_dev_reset_iommu_done() unless reset succeeds Nicolin Chen
2026-03-05 5:21 ` [PATCH v1 2/2] iommu/arm-smmu-v3: Recover ATC invalidate timeouts Nicolin Chen
2026-03-05 15:15 ` kernel test robot
2026-03-05 15:24 ` Robin Murphy
2026-03-05 21:06 ` Nicolin Chen
2026-03-05 23:30 ` Nicolin Chen
2026-03-05 23:52 ` Jason Gunthorpe
2026-03-06 15:24 ` Robin Murphy
2026-03-06 15:56 ` Jason Gunthorpe
2026-03-10 19:34 ` Pranjal Shrivastava
2026-03-05 15:39 ` Jason Gunthorpe
2026-03-05 21:15 ` Nicolin Chen
2026-03-05 23:41 ` Jason Gunthorpe
2026-03-06 1:29 ` Nicolin Chen
2026-03-06 1:33 ` Jason Gunthorpe
2026-03-06 5:06 ` Nicolin Chen
2026-03-06 13:02 ` Jason Gunthorpe
2026-03-06 19:20 ` Nicolin Chen
2026-03-06 19:22 ` Jason Gunthorpe
2026-03-06 19:39 ` Nicolin Chen
2026-03-06 19:47 ` Jason Gunthorpe
2026-03-10 19:40 ` Pranjal Shrivastava
2026-03-10 19:57 ` Nicolin Chen
2026-03-10 20:04 ` Pranjal Shrivastava
2026-03-06 13:22 ` Robin Murphy
2026-03-06 14:01 ` Jason Gunthorpe
2026-03-06 20:18 ` Nicolin Chen
2026-03-06 20:22 ` Jason Gunthorpe
2026-03-06 20:34 ` Nicolin Chen
2026-03-06 3:22 ` Baolu Lu
2026-03-06 13:00 ` Jason Gunthorpe
2026-03-06 19:35 ` Samiullah Khawaja
2026-03-06 19:43 ` Jason Gunthorpe
2026-03-06 19:59 ` Samiullah Khawaja
2026-03-06 20:03 ` Jason Gunthorpe
2026-03-06 20:22 ` Samiullah Khawaja
2026-03-06 20:26 ` Jason Gunthorpe
2026-03-10 20:00 ` Samiullah Khawaja
2026-03-11 12:12 ` Jason Gunthorpe
2026-03-06 2:35 ` kernel test robot
2026-03-10 19:16 ` Pranjal Shrivastava
2026-03-10 19:51 ` Nicolin Chen
2026-03-10 20:00 ` Pranjal Shrivastava
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=cover.1772686998.git.nicolinc@nvidia.com \
--to=nicolinc@nvidia.com \
--cc=Alexander.Grest@microsoft.com \
--cc=baolu.lu@linux.intel.com \
--cc=bhelgaas@google.com \
--cc=iommu@lists.linux.dev \
--cc=jgg@nvidia.com \
--cc=joro@8bytes.org \
--cc=kees@kernel.org \
--cc=kevin.tian@intel.com \
--cc=lenb@kernel.org \
--cc=linux-acpi@vger.kernel.org \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-pci@vger.kernel.org \
--cc=miko.lenczewski@arm.com \
--cc=praan@google.com \
--cc=rafael@kernel.org \
--cc=robin.murphy@arm.com \
--cc=smostafa@google.com \
--cc=vsethi@nvidia.com \
--cc=will@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox