[26.04_linux-nvidia-bos] ats always on v4 by nirmoy · Pull Request #401 · NVIDIA/NV-Kernels

nirmoy · 2026-04-30T09:17:11Z

LP: https://bugs.launchpad.net/ubuntu/+source/linux-nvidia-6.17/+bug/2150727

Summary:

Revert the older ATS always-on backport stack currently in the 26.04 BOS branch.
Backport v4 of the ATS always-on lore series:
- PCI support for CXL.cache-capable devices.
- PCI quirks for pre-CXL NVIDIA GPU and CX10 devices.
- arm-smmu-v3 support for devices that require ATS to stay enabled.
Keep NVIDIA SAUCE metadata, per-patch lore backport URLs, and Nirmoy backporter Signed-off-by trailers.

Backport notes:

Patch 1 is adapted to the 26.04 base where PCI_DVSEC_CXL_CACHE_CAPABLE already exists.
Patch 2 is applied after reverting the older ATS PCI quirk support.
Patch 3 is applied after reverting the older ATS arm-smmu-v3 support.

Validation:

Local patchscan: pass.
Local validate-pr: pass.
git diff --check: pass.
git show --check HEAD~6..HEAD: pass.
Targeted x86 object build: drivers/pci/ats.o, drivers/pci/quirks.o, drivers/pci/pci.o.
Tested on C1G2

github-actions · 2026-04-30T09:27:19Z

PR Validation Report

Patchscan ✅ No Missing Fixes

All cherry-picked commits checked — no missing upstream fixes found.

PR Lint ✅ All checks passed

Details

Checking 6 commits...

Cherry-pick digest:
┌──────────────┬──────────────────────────────────────────────────────────────────┬────────────┬─────────┬───────────────────────────┐
│ Local        │ Referenced upstream / Patch subject                              │ Patch-ID   │ Subject │ SoB chain                 │
├──────────────┼──────────────────────────────────────────────────────────────────┼────────────┼─────────┼───────────────────────────┤
│ 2287d554c55d │ iommu/arm-smmu-v3: allow ats to be always on                     │ noted      │ found   │ ok, backporter: nirmoyd   │
├──────────────┼──────────────────────────────────────────────────────────────────┼────────────┼─────────┼───────────────────────────┤
│ d4f429f68fb3 │ pci: allow ats to be always on for pre-cxl devices               │ noted      │ found   │ ok, backporter: nirmoyd   │
├──────────────┼──────────────────────────────────────────────────────────────────┼────────────┼─────────┼───────────────────────────┤
│ 5399b4322439 │ pci: allow ats to be always on for cxl.cache capable devices     │ noted      │ found   │ ok, backporter: nirmoyd   │
├──────────────┼──────────────────────────────────────────────────────────────────┼────────────┼─────────┼───────────────────────────┤
│ ab2033808093 │ [Revert] pci: allow ats to be always on for cxl.cache capable de │ N/A        │ N/A     │ nirmoyd                   │
├──────────────┼──────────────────────────────────────────────────────────────────┼────────────┼─────────┼───────────────────────────┤
│ 7323790a599c │ [Revert] pci: allow ats to be always on for non-cxl nvidia gpus  │ N/A        │ N/A     │ nirmoyd                   │
├──────────────┼──────────────────────────────────────────────────────────────────┼────────────┼─────────┼───────────────────────────┤
│ a4c077a94504 │ [Revert] iommu/arm-smmu-v3: allow ats to be always on            │ N/A        │ N/A     │ nirmoyd                   │
└──────────────┴──────────────────────────────────────────────────────────────────┴────────────┴─────────┴───────────────────────────┘

Lint: all checks passed.

BugLink: https://bugs.launchpad.net/ubuntu/+source/linux-nvidia-6.17/+bug/2150727 This reverts commit 11d6b22. Signed-off-by: Nirmoy Das <nirmoyd@nvidia.com>

… NVIDIA GPUs" BugLink: https://bugs.launchpad.net/ubuntu/+source/linux-nvidia-6.17/+bug/2150727 This reverts commit 208b48c. Signed-off-by: Nirmoy Das <nirmoyd@nvidia.com>

…he capable devices" BugLink: https://bugs.launchpad.net/ubuntu/+source/linux-nvidia-6.17/+bug/2150727 This reverts commit 967f9fe. Signed-off-by: Nirmoy Das <nirmoyd@nvidia.com>

…le devices BugLink: https://bugs.launchpad.net/ubuntu/+source/linux-nvidia-6.17/+bug/2150727 Controlled by the IOMMU driver, ATS is usually enabled "on demand" when a given PASID on a device is attached to an I/O page table. This is working even when a device has no translation on its RID (i.e., the RID is IOMMU bypassed). However, certain PCIe devices require non-PASID ATS on their RID even when the RID is IOMMU bypassed. Call this "always on". For example, CXL spec r4.0 notes in sec 3.2.5.13 Memory Type on CXL.cache: "To source requests on CXL.cache, devices need to get the Host Physical Address (HPA) from the Host by means of an ATS request on CXL.io." In other words, the CXL.cache capability requires ATS; otherwise, it can't access host physical memory. Introduce a new pci_ats_always_on() helper for the IOMMU driver to scan a PCI device and shift ATS policies between "on demand" and "always on". Add the support for CXL.cache devices first. Pre-CXL devices will be added in quirks.c file. Note that pci_ats_always_on() validates against pci_ats_supported(), so we ensure that untrusted devices (e.g. external ports) will not be always on. This maintains the existing ATS security policy regarding potential side- channel attacks via ATS. Cc: linux-cxl@vger.kernel.org Suggested-by: Vikram Sethi <vsethi@nvidia.com> Suggested-by: Jason Gunthorpe <jgg@nvidia.com> Reviewed-by: Jonathan Cameron <jonathan.cameron@huawei.com> Reviewed-by: Jason Gunthorpe <jgg@nvidia.com> Reviewed-by: Kevin Tian <kevin.tian@intel.com> Tested-by: Nirmoy Das <nirmoyd@nvidia.com> Acked-by: Nirmoy Das <nirmoyd@nvidia.com> Signed-off-by: Nicolin Chen <nicolinc@nvidia.com> Reviewed-by: Dave Jiang <dave.jiang@intel.com> (backported from https://lore.kernel.org/r/f6734b9dad0050138676f11ecd14e9db1cf6b697.1777269009.git.nicolinc@nvidia.com) [Nirmoy: Adapt to already existing PCI_DVSEC_CXL_CACHE_CAPABLE.] Signed-off-by: Nirmoy Das <nirmoyd@nvidia.com>

BugLink: https://bugs.launchpad.net/ubuntu/+source/linux-nvidia-6.17/+bug/2150727 Some NVIDIA GPU/NIC devices, though they don't implement CXL config space, have many CXL-like properties. Call this kind "pre-CXL". Similar to CXL.cache capability, these pre-CXL devices also require the ATS function even when their RIDs are IOMMU bypassed, i.e. keep ATS "always on" v.s. "on demand" when a non-zero PASID line gets enabled in SVA use cases. Introduce pci_dev_specific_ats_always_on() quirk function to scan a list of IDs for these devices. Then, include it in pci_ats_always_on(). Suggested-by: Jason Gunthorpe <jgg@nvidia.com> Reviewed-by: Nirmoy Das <nirmoyd@nvidia.com> Tested-by: Nirmoy Das <nirmoyd@nvidia.com> Reviewed-by: Jonathan Cameron <jonathan.cameron@huawei.com> Reviewed-by: Jason Gunthorpe <jgg@nvidia.com> Reviewed-by: Kevin Tian <kevin.tian@intel.com> Signed-off-by: Nicolin Chen <nicolinc@nvidia.com> Reviewed-by: Dave Jiang <dave.jiang@intel.com> (backported from https://lore.kernel.org/r/1a8cf5e88051ab5c10417edb94df598ecbc810cf.1777269009.git.nicolinc@nvidia.com) [Nirmoy: Apply after reverting older ATS always-on PCI quirk support.] Signed-off-by: Nirmoy Das <nirmoyd@nvidia.com>

BugLink: https://bugs.launchpad.net/ubuntu/+source/linux-nvidia-6.17/+bug/2150727 When a device's default substream attaches to an identity domain, the SMMU driver currently sets the device's STE between two modes: Mode 1: Cfg=Translate, S1DSS=Bypass, EATS=1 Mode 2: Cfg=bypass (EATS is ignored by HW) When there is an active PASID (non-default substream), mode 1 is used. And when there is no PASID support or no active PASID, mode 2 is used. The driver will also downgrade an STE from mode 1 to mode 2, when the last active substream becomes inactive. However, there are PCIe devices that demand ATS to be always on. For these devices, their STEs have to use the mode 1 as HW ignores EATS with mode 2. Change the driver accordingly: - always use the mode 1 - never downgrade to mode 2 - allocate and retain a CD table (see note below) Note that these devices might not support PASID, i.e. doing non-PASID ATS. In such a case, the ssid_bits is set to 0. However, s1cdmax must be set to a !0 value in order to keep the S1DSS field effective. Thus, when a master requires ats_always_on, set its s1cdmax to at least 1, meaning that the CD table will have a dummy entry (SSID=1) that will never be used. Now for these devices, arm_smmu_cdtab_allocated() will always return true, v.s. false prior to this change. When its default substream is attached to an IDENTITY domain, its first CD is NULL in the table, which is a totally valid case. Thus, add "!master->ats_always_on" to the condition. Reviewed-by: Jonathan Cameron <jonathan.cameron@huawei.com> Tested-by: Nirmoy Das <nirmoyd@nvidia.com> Acked-by: Nirmoy Das <nirmoyd@nvidia.com> Reviewed-by: Jason Gunthorpe <jgg@nvidia.com> Reviewed-by: Kevin Tian <kevin.tian@intel.com> Signed-off-by: Nicolin Chen <nicolinc@nvidia.com> Reviewed-by: Dave Jiang <dave.jiang@intel.com> (backported from https://lore.kernel.org/r/7403163ebf59380f88c7503b3adf0dae07428df8.1777269009.git.nicolinc@nvidia.com) [Nirmoy: Apply after reverting older ATS always-on arm-smmu-v3 support.] Signed-off-by: Nirmoy Das <nirmoyd@nvidia.com>

nvmochs · 2026-04-30T13:00:12Z

@nirmoy I see this received a RB tag from DaveJ. Do we need to take this in now or can we wait a week to see if it lands in -next?

nirmoy · 2026-04-30T14:24:11Z

@nirmoy I see this received a RB tag from DaveJ. Do we need to take this in now or can we wait a week to see if it lands in -next?

There is no rush, we can wait it to get merge in the -next

nirmoy · 2026-05-08T15:57:50Z

Checked the v4 Lore thread: no v5 appears expected, v4 has the review tags, but it is not in linux-next/pci/next/iommu/next yet, so I’ll keep this draft.

nirmoy · 2026-05-12T17:40:42Z

nvidia@localhost:/home/nvidia/nirmoy/linux$ sudo dmesg|grep smmu -i |grep event -i
nvidia@localhost:/home/nvidia/nirmoy/linux$ sudo dmesg|grep cache_capable
[    7.425771] pci 0002:81:00.0: CXL DVSEC cap 0x8bd7 cache_capable=1 <--- I added this to debug the kernel
[    7.426036] pci 0002:c1:00.0: CXL DVSEC cap 0x8bd7 cache_capable=1
nvidia@localhost:/home/nvidia/nirmoy/linux$ sudo nvidia-smi
Tue May 12 17:40:07 2026
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 610.26                 KMD Version: 610.26        CUDA UMD Version: 13.3     |
+-----------------------------------------+------------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
|                                         |                        |               MIG M. |
|=========================================+========================+======================|
|   0  NVIDIA Graphics Device         Off |   00000002:81:00.0 Off |                    0 |
| N/A   28C    P5            362W / 2300W |       0MiB / 286524MiB |      0%      Default |
|                                         |                        |             Disabled |
+-----------------------------------------+------------------------+----------------------+
|   1  NVIDIA Graphics Device         Off |   00000002:C1:00.0 Off |                    0 |
| N/A   28C    P5            357W / 2300W |       0MiB / 286524MiB |      0%      Default |
|                                         |                        |             Disabled |
+-----------------------------------------+------------------------+----------------------+

+-----------------------------------------------------------------------------------------+
| Processes:                                                                              |
|  GPU   GI   CI              PID   Type   Process name                        GPU Memory |
|        ID   ID                                                               Usage      |
|=========================================================================================|
|  No running processes found                                                             |
+-----------------------------------------------------------------------------------------+

clsotog

Acked-by: Carol L Soto <csoto@nvidia.com>

jamieNguyenNVIDIA · 2026-05-12T18:55:37Z

@nirmoy: The change looks good, but I have a nitpick -- the convention is to have your backport note go after the "(backported from...)" information.

(backported from https://lore.kernel.org/r/7403163ebf59380f88c7503b3adf0dae07428df8.1777269009.git.nicolinc@nvidia.com)
[Nirmoy: Apply after reverting older ATS always-on arm-smmu-v3 support.]
Signed-off-by: Nirmoy Das <nirmoyd@nvidia.com>

Instead of:

[Nirmoy: Apply after reverting older ATS always-on arm-smmu-v3 support.]
(backported from https://lore.kernel.org/r/7403163ebf59380f88c7503b3adf0dae07428df8.1777269009.git.nicolinc@nvidia.com)
Signed-off-by: Nirmoy Das <nirmoyd@nvidia.com>

nvmochs

I reviewed the PR with codex and did not find anything of significance.

With the nit that Jamie pointed out resolved:
Acked-by: Matthew R. Ochs <mochs@nvidia.com>

nirmoy · 2026-05-13T12:55:03Z

Thanks @jamieNguyenNVIDIA Added that note in the PR validation script so it capture next time.

nirmoy marked this pull request as draft April 30, 2026 09:26

nirmoy changed the title ~~My ats always on v4~~ ats always on v4 Apr 30, 2026

nirmoy force-pushed the my-ats-always-on-v4 branch from ed22fc3 to 072939e Compare April 30, 2026 09:36

nirmoy changed the title ~~ats always on v4~~ [26.04_linux-nvidia-bos] ats always on v4 Apr 30, 2026

nirmoy added 3 commits April 30, 2026 02:44

Revert "NVIDIA: VR: SAUCE: iommu/arm-smmu-v3: Allow ATS to be always on"

a4c077a

BugLink: https://bugs.launchpad.net/ubuntu/+source/linux-nvidia-6.17/+bug/2150727 This reverts commit 11d6b22. Signed-off-by: Nirmoy Das <nirmoyd@nvidia.com>

Revert "NVIDIA: VR: SAUCE: PCI: Allow ATS to be always on for non-CXL…

7323790

… NVIDIA GPUs" BugLink: https://bugs.launchpad.net/ubuntu/+source/linux-nvidia-6.17/+bug/2150727 This reverts commit 208b48c. Signed-off-by: Nirmoy Das <nirmoyd@nvidia.com>

Revert "NVIDIA: VR: SAUCE: PCI: Allow ATS to be always on for CXL.cac…

ab20338

…he capable devices" BugLink: https://bugs.launchpad.net/ubuntu/+source/linux-nvidia-6.17/+bug/2150727 This reverts commit 967f9fe. Signed-off-by: Nirmoy Das <nirmoyd@nvidia.com>

nirmoy force-pushed the my-ats-always-on-v4 branch 2 times, most recently from 15c3560 to 520182f Compare April 30, 2026 09:53

nicolinc added 3 commits April 30, 2026 03:03

nirmoy force-pushed the my-ats-always-on-v4 branch from 520182f to 80c6442 Compare April 30, 2026 10:04

nirmoy marked this pull request as ready for review May 12, 2026 17:40

nirmoy requested review from clsotog, jamieNguyenNVIDIA and nvmochs May 12, 2026 17:55

clsotog approved these changes May 12, 2026

View reviewed changes

jamieNguyenNVIDIA approved these changes May 12, 2026

View reviewed changes

nvmochs approved these changes May 12, 2026

View reviewed changes

This was referenced May 13, 2026

validate-pr: enforce backport note trailer order nirmoy/NV-Kernels#15

Merged

TEST: clone NVIDIA PR #401 for backport-note validation nirmoy/NV-Kernels#16

Open

validate-pr: enforce backport note trailer order #418

Open

nirmoy force-pushed the my-ats-always-on-v4 branch from 80c6442 to 2287d55 Compare May 13, 2026 12:56

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[26.04_linux-nvidia-bos] ats always on v4#401

[26.04_linux-nvidia-bos] ats always on v4#401
nirmoy wants to merge 6 commits into
NVIDIA:26.04_linux-nvidia-bosfrom
nirmoy:my-ats-always-on-v4

nirmoy commented Apr 30, 2026 •

edited

Loading

Uh oh!

github-actions Bot commented Apr 30, 2026 •

edited

Loading

Uh oh!

nvmochs commented Apr 30, 2026

Uh oh!

nirmoy commented Apr 30, 2026

Uh oh!

nirmoy commented May 8, 2026

Uh oh!

nirmoy commented May 12, 2026 •

edited

Loading

Uh oh!

clsotog left a comment

Uh oh!

jamieNguyenNVIDIA commented May 12, 2026 •

edited

Loading

Uh oh!

nvmochs left a comment

Uh oh!

nirmoy commented May 13, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

Conversation

nirmoy commented Apr 30, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

github-actions Bot commented Apr 30, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

PR Validation Report

Patchscan ✅ No Missing Fixes

PR Lint ✅ All checks passed

Uh oh!

nvmochs commented Apr 30, 2026

Uh oh!

nirmoy commented Apr 30, 2026

Uh oh!

nirmoy commented May 8, 2026

Uh oh!

nirmoy commented May 12, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

clsotog left a comment

Choose a reason for hiding this comment

Uh oh!

jamieNguyenNVIDIA commented May 12, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

nvmochs left a comment

Choose a reason for hiding this comment

Uh oh!

nirmoy commented May 13, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

nirmoy commented Apr 30, 2026 •

edited

Loading

github-actions Bot commented Apr 30, 2026 •

edited

Loading

nirmoy commented May 12, 2026 •

edited

Loading

jamieNguyenNVIDIA commented May 12, 2026 •

edited

Loading