fix: [linux-nvidia-6.6] Resolve IOMMU passthrough conflicts with NVIDIA UVM on ARM64 causing NEMO/PyTorch failures

.github/workflows/kernel-build.yml

@@ -0,0 +1,77 @@
+name: Kernel Test
+
+on:
+  push:
+  pull_request:
+
+jobs:
+  build:
+    name: ${{ matrix.runs-on == 'ubuntu-latest' && 'x86_64' || 'ARM64' }} - ${{ matrix.config }}
+    runs-on: ${{ matrix.runs-on }}
+    strategy:
+      matrix:
+        include:
+          - runs-on: ubuntu-latest
+            config: defconfig
+          - runs-on: ubuntu-24.04-arm
+            config: config_4k
+          - runs-on: ubuntu-24.04-arm
+            config: config_64k
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+
+      - name: Install Dependencies
+        run: |
+          sudo apt-get update -y
+          sudo apt-get install -y build-essential flex bison libssl-dev bc \
+            busybox-static cmake coreutils virtme-ng cpio elfutils file gcc \
+            git iproute2 jq kbd kmod libcap-dev libelf-dev libunwind-dev \
+            libvirt-clients libzstd-dev libasound2-dev libdrm-dev \
+            linux-headers-generic linux-tools-common linux-tools-generic \
+            make ninja-build pahole pkg-config python3-dev python3-pip \
+            python3-requests qemu-kvm qemu-system-arm rsync stress-ng udev zstd \
+            libseccomp-dev libcap-ng-dev llvm-19 clang-19 python3-full curl \
+            bpftrace dwarves gcc-aarch64-linux-gnu binutils-aarch64-linux-gnu \
+            bash-completion ccache debhelper devscripts docbook-utils \
+            fakeroot gawk kernel-wedge libncurses5-dev makedumpfile python3-git \
+            python3-requests python3-ruamel.yaml sharutils transfig ubuntu-dev-tools \
+            libnuma-dev xmlto
+
+      - name: Install virtme-ng
+        run: |
+          git clone https://github.com/arighi/virtme-ng.git
+          cd virtme-ng
+          VNG_PACKAGE=1 BUILD_VIRTME_NG_INIT=1 pip install --break-system-packages .
+          cd -
+
+      - name: Configure Kernel
+        run: |
+          if [ "${{ matrix.config }}" = "defconfig" ]; then
+            vng -vb --kconfig
+          else
+            vng -vb --kconfig --config .github/workflows/${{ matrix.config }}
+          fi
+          # Ensure WERROR is disabled
+          ./scripts/config --disable CONFIG_WERROR
+          # Disable certificate verification
+          ./scripts/config --disable SYSTEM_TRUSTED_KEYS
+          ./scripts/config --disable SYSTEM_REVOCATION_KEYS ""
+
+      - name: Build Kernel
+        run: |
+          vng -vb
+
+      - name: Boot test
+        run: |
+          vng -v -m 4G -- uname -a
+
+      - name: Run kselftests
+        run: |
+          make headers_install
+          make -j$(nproc) -C tools/testing/selftests \
+            TARGETS="mm dma iommu locking" \
+            INSTALL_PATH=$(pwd)/tests \
+            all install
+          vng -v -m 4G -- ./tests/run_kselftest.sh --summary

.gitignore

@@ -11,6 +11,7 @@
 # Normal rules (sorted alphabetically)
 #
 .*
+!.github/
 *.a
 *.asn1.[ch]
 *.bin

Documentation/ABI/testing/sysfs-bus-coresight-devices-tpdm

@@ -11,3 +11,162 @@ Description:
 		Accepts only one of the 2 values -  1 or 2.
 		1 : Generate 64 bits data
 		2 : Generate 32 bits data
+
+What:		/sys/bus/coresight/devices/<tpdm-name>/reset_dataset
+Date:		March 2023
+KernelVersion	6.7
+Contact:	Jinlong Mao (QUIC) <quic_jinlmao@quicinc.com>, Tao Zhang (QUIC) <quic_taozha@quicinc.com>
+Description:
+		(Write) Reset the dataset of the tpdm.
+
+		Accepts only one value -  1.
+		1 : Reset the dataset of the tpdm
+
+What:		/sys/bus/coresight/devices/<tpdm-name>/dsb_trig_type
+Date:		March 2023
+KernelVersion	6.7
+Contact:	Jinlong Mao (QUIC) <quic_jinlmao@quicinc.com>, Tao Zhang (QUIC) <quic_taozha@quicinc.com>
+Description:
+		(RW) Set/Get the trigger type of the DSB for tpdm.
+
+		Accepts only one of the 2 values -  0 or 1.
+		0 : Set the DSB trigger type to false
+		1 : Set the DSB trigger type to true
+
+What:		/sys/bus/coresight/devices/<tpdm-name>/dsb_trig_ts
+Date:		March 2023
+KernelVersion	6.7
+Contact:	Jinlong Mao (QUIC) <quic_jinlmao@quicinc.com>, Tao Zhang (QUIC) <quic_taozha@quicinc.com>
+Description:
+		(RW) Set/Get the trigger timestamp of the DSB for tpdm.
+
+		Accepts only one of the 2 values -  0 or 1.
+		0 : Set the DSB trigger type to false
+		1 : Set the DSB trigger type to true
+
+What:		/sys/bus/coresight/devices/<tpdm-name>/dsb_mode
+Date:		March 2023
+KernelVersion	6.7
+Contact:	Jinlong Mao (QUIC) <quic_jinlmao@quicinc.com>, Tao Zhang (QUIC) <quic_taozha@quicinc.com>
+Description:
+		(RW) Set/Get the programming mode of the DSB for tpdm.
+
+		Accepts the value needs to be greater than 0. What data
+		bits do is listed below.
+		Bit[0:1] : Test mode control bit for choosing the inputs.
+		Bit[3] : Set to 0 for low performance mode.
+				 Set to 1 for high performance mode.
+		Bit[4:8] : Select byte lane for high performance mode.
+
+What:		/sys/bus/coresight/devices/<tpdm-name>/dsb_edge/ctrl_idx
+Date:		March 2023
+KernelVersion	6.7
+Contact:	Jinlong Mao (QUIC) <quic_jinlmao@quicinc.com>, Tao Zhang (QUIC) <quic_taozha@quicinc.com>
+Description:
+		(RW) Set/Get the index number of the edge detection for the DSB
+		subunit TPDM. Since there are at most 256 edge detections, this
+		value ranges from 0 to 255.
+
+What:		/sys/bus/coresight/devices/<tpdm-name>/dsb_edge/ctrl_val
+Date:		March 2023
+KernelVersion	6.7
+Contact:	Jinlong Mao (QUIC) <quic_jinlmao@quicinc.com>, Tao Zhang (QUIC) <quic_taozha@quicinc.com>
+Description:
+		Write a data to control the edge detection corresponding to
+		the index number. Before writing data to this sysfs file,
+		"ctrl_idx" should be written first to configure the index
+		number of the edge detection which needs to be controlled.
+
+		Accepts only one of the following values.
+		0 - Rising edge detection
+		1 - Falling edge detection
+		2 - Rising and falling edge detection (toggle detection)
+
+
+What:		/sys/bus/coresight/devices/<tpdm-name>/dsb_edge/ctrl_mask
+Date:		March 2023
+KernelVersion	6.7
+Contact:	Jinlong Mao (QUIC) <quic_jinlmao@quicinc.com>, Tao Zhang (QUIC) <quic_taozha@quicinc.com>
+Description:
+		Write a data to mask the edge detection corresponding to the index
+		number. Before writing data to this sysfs file, "ctrl_idx" should
+		be written first to configure the index number of the edge detection
+		which needs to be masked.
+
+		Accepts only one of the 2 values -  0 or 1.
+
+What:		/sys/bus/coresight/devices/<tpdm-name>/dsb_edge/edcr[0:15]
+Date:		March 2023
+KernelVersion	6.7
+Contact:	Jinlong Mao (QUIC) <quic_jinlmao@quicinc.com>, Tao Zhang (QUIC) <quic_taozha@quicinc.com>
+Description:
+		Read a set of the edge control value of the DSB in TPDM.
+
+What:		/sys/bus/coresight/devices/<tpdm-name>/dsb_edge/edcmr[0:7]
+Date:		March 2023
+KernelVersion	6.7
+Contact:	Jinlong Mao (QUIC) <quic_jinlmao@quicinc.com>, Tao Zhang (QUIC) <quic_taozha@quicinc.com>
+Description:
+		Read a set of the edge control mask of the DSB in TPDM.
+
+What:		/sys/bus/coresight/devices/<tpdm-name>/dsb_trig_patt/xpr[0:7]
+Date:		March 2023
+KernelVersion	6.7
+Contact:	Jinlong Mao (QUIC) <quic_jinlmao@quicinc.com>, Tao Zhang (QUIC) <quic_taozha@quicinc.com>
+Description:
+		(RW) Set/Get the value of the trigger pattern for the DSB
+		subunit TPDM.
+
+What:		/sys/bus/coresight/devices/<tpdm-name>/dsb_trig_patt/xpmr[0:7]
+Date:		March 2023
+KernelVersion	6.7
+Contact:	Jinlong Mao (QUIC) <quic_jinlmao@quicinc.com>, Tao Zhang (QUIC) <quic_taozha@quicinc.com>
+Description:
+		(RW) Set/Get the mask of the trigger pattern for the DSB
+		subunit TPDM.
+
+What:		/sys/bus/coresight/devices/<tpdm-name>/dsb_patt/tpr[0:7]
+Date:		March 2023
+KernelVersion	6.7
+Contact:	Jinlong Mao (QUIC) <quic_jinlmao@quicinc.com>, Tao Zhang (QUIC) <quic_taozha@quicinc.com>
+Description:
+		(RW) Set/Get the value of the pattern for the DSB subunit TPDM.
+
+What:		/sys/bus/coresight/devices/<tpdm-name>/dsb_patt/tpmr[0:7]
+Date:		March 2023
+KernelVersion	6.7
+Contact:	Jinlong Mao (QUIC) <quic_jinlmao@quicinc.com>, Tao Zhang (QUIC) <quic_taozha@quicinc.com>
+Description:
+		(RW) Set/Get the mask of the pattern for the DSB subunit TPDM.
+
+What:		/sys/bus/coresight/devices/<tpdm-name>/dsb_patt/enable_ts
+Date:		March 2023
+KernelVersion	6.7
+Contact:	Jinlong Mao (QUIC) <quic_jinlmao@quicinc.com>, Tao Zhang (QUIC) <quic_taozha@quicinc.com>
+Description:
+		(Write) Set the pattern timestamp of DSB tpdm. Read
+		the pattern timestamp of DSB tpdm.
+
+		Accepts only one of the 2 values -  0 or 1.
+		0 : Disable DSB pattern timestamp.
+		1 : Enable DSB pattern timestamp.
+
+What:		/sys/bus/coresight/devices/<tpdm-name>/dsb_patt/set_type
+Date:		March 2023
+KernelVersion	6.7
+Contact:	Jinlong Mao (QUIC) <quic_jinlmao@quicinc.com>, Tao Zhang (QUIC) <quic_taozha@quicinc.com>
+Description:
+		(Write) Set the pattern type of DSB tpdm. Read
+		the pattern type of DSB tpdm.
+
+		Accepts only one of the 2 values -  0 or 1.
+		0 : Set the DSB pattern type to value.
+		1 : Set the DSB pattern type to toggle.
+
+What:		/sys/bus/coresight/devices/<tpdm-name>/dsb_msr/msr[0:31]
+Date:		March 2023
+KernelVersion	6.7
+Contact:	Jinlong Mao (QUIC) <quic_jinlmao@quicinc.com>, Tao Zhang (QUIC) <quic_taozha@quicinc.com>
+Description:
+		(RW) Set/Get the MSR(mux select register) for the DSB subunit
+		TPDM.

Documentation/admin-guide/kernel-parameters.txt

@@ -4496,6 +4496,38 @@
 				bridges without forcing it upstream. Note:
 				this removes isolation between devices and
 				may put more devices in an IOMMU group.
+		config_acs=
+				Format:
+				<ACS flags>@<pci_dev>[; ...]
+				Specify one or more PCI devices (in the format
+				specified above) optionally prepended with flags
+				and separated by semicolons. The respective
+				capabilities will be enabled, disabled or
+				unchanged based on what is specified in
+				flags.
+
+				ACS Flags is defined as follows:
+				  bit-0 : ACS Source Validation
+				  bit-1 : ACS Translation Blocking
+				  bit-2 : ACS P2P Request Redirect
+				  bit-3 : ACS P2P Completion Redirect
+				  bit-4 : ACS Upstream Forwarding
+				  bit-5 : ACS P2P Egress Control
+				  bit-6 : ACS Direct Translated P2P
+				Each bit can be marked as:
+				  '0' – force disabled
+				  '1' – force enabled
+				  'x' – unchanged
+				For example,
+				  pci=config_acs=10x
+				would configure all devices that support
+				ACS to enable P2P Request Redirect, disable
+				Translation Blocking, and leave Source
+				Validation unchanged from whatever power-up
+				or firmware set it to.
+
+				Note: this may remove isolation between devices
+				and may put more devices in an IOMMU group.
 		force_floating	[S390] Force usage of floating interrupts.
 		nomio		[S390] Do not use MIO instructions.
 		norid		[S390] ignore the RID field and force use of