linux-stable.git/tools/testing/selftests/vfio, branch linux-rolling-stable Linux kernel stable tree vfio: selftests: Ensure libvfio output dirs are always created 2026-07-24T14:19:51+00:00 David Matlack dmatlack@google.com 2026-06-11T21:39:45+00:00 93299544227948cab8b2700aa7198075278f677b [ Upstream commit 785562e31dbcd85ca583cf58c446e63aa8a5af08 ] Add an explicit dependency between the output object files and the output directories that need to be created to hold those files. This ensures that the output directories are always created. Creating the output directories at parse time (current behavior) doesn't support the scenario where someone does "make clean all". The directories will be created during parsing, deleted during "clean" and then not available for the "all" target. Use an order-only prerequisite for the output directories, rather than a normal prerequisite, to avoid unnecessary recompilations. Fixes: 19faf6fd969c ("vfio: selftests: Add a helper library for VFIO selftests") Reported-by: Sashiko <sashiko-bot@kernel.org> Closes: https://lore.kernel.org/kvm/20260610010314.DB8861F00893@smtp.kernel.org/ Signed-off-by: David Matlack <dmatlack@google.com> Link: https://lore.kernel.org/r/20260611213945.3714421-1-dmatlack@google.com Signed-off-by: Alex Williamson <alex@shazbot.org> Signed-off-by: Sasha Levin <sashal@kernel.org> 
[ Upstream commit 785562e31dbcd85ca583cf58c446e63aa8a5af08 ]

Add an explicit dependency between the output object files and the
output directories that need to be created to hold those files. This
ensures that the output directories are always created.

Creating the output directories at parse time (current behavior) doesn't
support the scenario where someone does "make clean all". The
directories will be created during parsing, deleted during "clean" and
then not available for the "all" target.

Use an order-only prerequisite for the output directories, rather than a
normal prerequisite, to avoid unnecessary recompilations.

Fixes: 19faf6fd969c ("vfio: selftests: Add a helper library for VFIO selftests")
Reported-by: Sashiko <sashiko-bot@kernel.org>
Closes: https://lore.kernel.org/kvm/20260610010314.DB8861F00893@smtp.kernel.org/
Signed-off-by: David Matlack <dmatlack@google.com>
Link: https://lore.kernel.org/r/20260611213945.3714421-1-dmatlack@google.com
Signed-off-by: Alex Williamson <alex@shazbot.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
vfio: selftests: Allow builds when ARCH=x86 2026-07-24T14:19:16+00:00 David Matlack dmatlack@google.com 2026-04-28T23:27:06+00:00 1203f9ea8fd4a343f9e27b5be9d1d2075b244fc5 [ Upstream commit 75182529687e204b2bd046a3f7b1c2acb5031032 ] Allow builds when ARCH=x86 since the top-level Makefile can set ARCH=x86 even for 64-bit x86 builds. Note that ARCH=x86 could also indicate a native build on a 32-bit x86 host. However, it doesn't seem like anyone is building selftests natively on 32-bit x86 hosts these days since KVM selftests allow ARCH=x86 and fail to compile on 32-bit x86. If someone reports an issue on 32-bit native builds we can harden the KVM and VFIO selftests to explicitly check 64-bit (see the discussion in the Closes link below). Fixes: a55d4bbbe644 ("vfio: selftests: only build tests on arm64 and x86_64") Reported-by: Jason Gunthorpe <jgg@nvidia.com> Closes: https://lore.kernel.org/kvm/20260427231217.GA1670652@nvidia.com/ Signed-off-by: David Matlack <dmatlack@google.com> Reviewed-by: Jason Gunthorpe <jgg@nvidia.com> Link: https://lore.kernel.org/r/20260428232707.2139059-1-dmatlack@google.com Signed-off-by: Alex Williamson <alex@shazbot.org> Signed-off-by: Sasha Levin <sashal@kernel.org> 
[ Upstream commit 75182529687e204b2bd046a3f7b1c2acb5031032 ]

Allow builds when ARCH=x86 since the top-level Makefile can set ARCH=x86
even for 64-bit x86 builds.

Note that ARCH=x86 could also indicate a native build on a 32-bit x86
host. However, it doesn't seem like anyone is building selftests
natively on 32-bit x86 hosts these days since KVM selftests allow
ARCH=x86 and fail to compile on 32-bit x86.

If someone reports an issue on 32-bit native builds we can harden the
KVM and VFIO selftests to explicitly check 64-bit (see the discussion in
the Closes link below).

Fixes: a55d4bbbe644 ("vfio: selftests: only build tests on arm64 and x86_64")
Reported-by: Jason Gunthorpe <jgg@nvidia.com>
Closes: https://lore.kernel.org/kvm/20260427231217.GA1670652@nvidia.com/
Signed-off-by: David Matlack <dmatlack@google.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Link: https://lore.kernel.org/r/20260428232707.2139059-1-dmatlack@google.com
Signed-off-by: Alex Williamson <alex@shazbot.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
vfio: selftests: Fix out-of-tree build with make O= 2026-07-24T14:19:16+00:00 Jason Gunthorpe jgg@nvidia.com 2026-05-14T16:04:44+00:00 8d33d107c1308ac02c4e0e1331852cb951667516 [ Upstream commit a21b864bd951e452922dbd66747da24daca2b04f ] The test programs are compiled via a static pattern rule that requires intermediate .o files: $(TEST_GEN_PROGS): %: %.o $(LIBVFIO_O) After lib.mk prefixes TEST_GEN_PROGS with $(OUTPUT), this creates dependencies on .o files in the output directory (e.g. $(OUTPUT)/vfio_dma_mapping_test.o). However, there is no rule to compile these .o files from the source directory .c files when OUTPUT differs from the source directory. Add an explicit chain of pattern rules: $(OUTPUT)/% -> $(OUTPUT)/%.o -> %.c Following the same pattern already used in libvfio.mk for the library objects. Fixes: 19faf6fd969c ("vfio: selftests: Add a helper library for VFIO selftests") Reviewed-by: David Matlack <dmatlack@google.com> Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> Link: https://lore.kernel.org/r/0-v2-4ccc247e6aff+1d93-vfio_st_make_o_jgg@nvidia.com Signed-off-by: Alex Williamson <alex@shazbot.org> Signed-off-by: Sasha Levin <sashal@kernel.org> 
[ Upstream commit a21b864bd951e452922dbd66747da24daca2b04f ]

The test programs are compiled via a static pattern rule that requires
intermediate .o files:

  $(TEST_GEN_PROGS): %: %.o $(LIBVFIO_O)

After lib.mk prefixes TEST_GEN_PROGS with $(OUTPUT), this creates
dependencies on .o files in the output directory (e.g.
$(OUTPUT)/vfio_dma_mapping_test.o). However, there is no rule to compile
these .o files from the source directory .c files when OUTPUT differs
from the source directory.

Add an explicit chain of pattern rules:
  $(OUTPUT)/% -> $(OUTPUT)/%.o -> %.c

Following the same pattern already used in libvfio.mk for the library
objects.

Fixes: 19faf6fd969c ("vfio: selftests: Add a helper library for VFIO selftests")
Reviewed-by: David Matlack <dmatlack@google.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Link: https://lore.kernel.org/r/0-v2-4ccc247e6aff+1d93-vfio_st_make_o_jgg@nvidia.com
Signed-off-by: Alex Williamson <alex@shazbot.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
vfio: selftests: Support DMR and GNR-D DSA devices 2026-03-20T19:54:51+00:00 Yi Lai yi1.lai@intel.com 2026-03-20T01:09:29+00:00 c82cfe15916d33e89c2d2efeeb624e8c9c2c4ca8 Currently, the VFIO DSA driver test only supports the SPR DSA device ID. Attempting to run the test on newer platforms like DMR or GNR-D results in a "No driver found" error, causing the test to be skipped. Refactor dsa_probe() to use a switch statement for checking device IDs. This improves maintainability and makes it easier to add new device IDs in the future. Add the following DSA device IDs to the supported list: PCI_DEVICE_ID_INTEL_DSA_DMR (0x1212) PCI_DEVICE_ID_INTEL_DSA_GNRD (0x11fb) Signed-off-by: Yi Lai <yi1.lai@intel.com> Reviewed-by: David Matlack <dmatlack@google.com> Link: https://lore.kernel.org/r/20260320010930.481380-1-yi1.lai@intel.com Signed-off-by: Alex Williamson <alex@shazbot.org> 
Currently, the VFIO DSA driver test only supports the SPR DSA device ID.
Attempting to run the test on newer platforms like DMR or GNR-D results
in a "No driver found" error, causing the test to be skipped.

Refactor dsa_probe() to use a switch statement for checking device IDs.
This improves maintainability and makes it easier to add new device IDs
in the future.

Add the following DSA device IDs to the supported list:
PCI_DEVICE_ID_INTEL_DSA_DMR  (0x1212)
PCI_DEVICE_ID_INTEL_DSA_GNRD (0x11fb)

Signed-off-by: Yi Lai <yi1.lai@intel.com>
Reviewed-by: David Matlack <dmatlack@google.com>
Link: https://lore.kernel.org/r/20260320010930.481380-1-yi1.lai@intel.com
Signed-off-by: Alex Williamson <alex@shazbot.org>
vfio: selftests: Build tests on aarch64 2026-03-20T19:52:16+00:00 Ted Logan tedlogan@fb.com 2026-03-19T22:58:47+00:00 1347a742a1e1b080e2e8d200312ae45b8d6ac859 Fix vfio selftests on aarch64, allowing native builds on aarch64 hosts. Reported-by: Matt Evans <mattev@meta.com> Closes: https://lore.kernel.org/all/e51b4ff2-13c4-47d4-b781-3dcbd740d274@meta.com/ Fixes: a55d4bbbe644 ("vfio: selftests: only build tests on arm64 and x86_64") Signed-off-by: Ted Logan <tedlogan@fb.com> Reviewed-by: David Matlack <dmatlack@google.com> Link: https://lore.kernel.org/r/20260319-vfio-selftests-aarch64-v2-1-bb2621c24dc4@fb.com Signed-off-by: Alex Williamson <alex@shazbot.org> 
Fix vfio selftests on aarch64, allowing native builds on aarch64 hosts.

Reported-by: Matt Evans <mattev@meta.com>
Closes: https://lore.kernel.org/all/e51b4ff2-13c4-47d4-b781-3dcbd740d274@meta.com/
Fixes: a55d4bbbe644 ("vfio: selftests: only build tests on arm64 and x86_64")
Signed-off-by: Ted Logan <tedlogan@fb.com>
Reviewed-by: David Matlack <dmatlack@google.com>
Link: https://lore.kernel.org/r/20260319-vfio-selftests-aarch64-v2-1-bb2621c24dc4@fb.com
Signed-off-by: Alex Williamson <alex@shazbot.org>
vfio: selftests: Fix VLA initialisation in vfio_pci_irq_set() 2026-03-19T18:32:08+00:00 Manish Honap mhonap@nvidia.com 2026-03-17T05:14:02+00:00 4f42d716707654134789a0205a050b0d022be948 C does not permit an initialiser expression on a variable-length array (C99 Section 6.7.9 constraint: "The type of the entity to be initialized shall not be a variable length array type"). vfio_pci_irq_set() declared: u8 buf[sizeof(struct vfio_irq_set) + sizeof(int) * count] = {}; where `count` is a runtime function parameter, making `buf` a VLA. GCC rejects this with (tried with GCC-9.4.0): error: variable-sized object may not be initialized Fix by removing the `= {}` initialiser and inserting an explicit memset() immediately after the declaration. memset() on a VLA is perfectly legal and achieves the same zero-initialisation on all conforming C implementations. Fixes: 19faf6fd969c ("vfio: selftests: Add a helper library for VFIO selftests") Cc: stable@vger.kernel.org Reviewed-by: Dave Jiang <dave.jiang@intel.com> Reviewed-by: David Matlack <dmatlack@google.com> Signed-off-by: Manish Honap <mhonap@nvidia.com> Link: https://lore.kernel.org/r/20260317051402.3725670-1-mhonap@nvidia.com Signed-off-by: Alex Williamson <alex@shazbot.org> 
C does not permit an initialiser expression on a variable-length array
(C99 Section 6.7.9 constraint: "The type of the entity to be initialized
shall not be a variable length array type").

vfio_pci_irq_set() declared:

      u8 buf[sizeof(struct vfio_irq_set) + sizeof(int) * count] = {};

where `count` is a runtime function parameter, making `buf` a VLA.

GCC rejects this with (tried with GCC-9.4.0):

      error: variable-sized object may not be initialized

Fix by removing the `= {}` initialiser and inserting an explicit
memset() immediately after the declaration.  memset() on a VLA is
perfectly legal and achieves the same zero-initialisation on all
conforming C implementations.

Fixes: 19faf6fd969c ("vfio: selftests: Add a helper library for VFIO selftests")
Cc: stable@vger.kernel.org
Reviewed-by: Dave Jiang <dave.jiang@intel.com>
Reviewed-by: David Matlack <dmatlack@google.com>
Signed-off-by: Manish Honap <mhonap@nvidia.com>
Link: https://lore.kernel.org/r/20260317051402.3725670-1-mhonap@nvidia.com
Signed-off-by: Alex Williamson <alex@shazbot.org>
vfio: selftests: fix crash in vfio_dma_mapping_mmio_test 2026-03-10T17:56:01+00:00 Alex Mastro amastro@fb.com 2026-03-03T19:46:24+00:00 f183963891b4b0126f19aa0993ed931f3f3f9520 Remove the __iommu_unmap() call on a region that was never mapped. When __iommu_map() fails (expected for MMIO vaddrs in non-VFIO modes), the region is not added to the dma_regions list, leaving its list_head zero-initialized. If the unmap ioctl returns success, __iommu_unmap() calls list_del_init() on this zeroed node and crashes. This fixes the iommufd_compat_type1 and iommufd_compat_type1v2 test variants. Fixes: 080723f4d4c3 ("vfio: selftests: Add vfio_dma_mapping_mmio_test") Signed-off-by: Alex Mastro <amastro@fb.com> Reviewed-by: David Matlack <dmatlack@google.com> Reviewed-by: Yuan Yao <yaoyuan@linux.alibaba.com> Link: https://lore.kernel.org/r/20260303-fix-mmio-test-v1-1-78b4a9e46a4e@fb.com Signed-off-by: Alex Williamson <alex@shazbot.org> 
Remove the __iommu_unmap() call on a region that was never mapped.
When __iommu_map() fails (expected for MMIO vaddrs in non-VFIO
modes), the region is not added to the dma_regions list, leaving its
list_head zero-initialized. If the unmap ioctl returns success,
__iommu_unmap() calls list_del_init() on this zeroed node and crashes.

This fixes the iommufd_compat_type1 and iommufd_compat_type1v2
test variants.

Fixes: 080723f4d4c3 ("vfio: selftests: Add vfio_dma_mapping_mmio_test")
Signed-off-by: Alex Mastro <amastro@fb.com>
Reviewed-by: David Matlack <dmatlack@google.com>
Reviewed-by: Yuan Yao <yaoyuan@linux.alibaba.com>
Link: https://lore.kernel.org/r/20260303-fix-mmio-test-v1-1-78b4a9e46a4e@fb.com
Signed-off-by: Alex Williamson <alex@shazbot.org>
vfio: selftests: only build tests on arm64 and x86_64 2026-02-06T22:08:05+00:00 Ted Logan tedlogan@fb.com 2026-02-03T01:23:53+00:00 a55d4bbbe64494bb92b32402018efb2ffc44d796 Only build vfio self-tests on arm64 and x86_64; these are the only architectures where the vfio self-tests are run. Addresses compiler warnings for format and conversions on i386. Reported-by: kernel test robot <lkp@intel.com> Closes: https://lore.kernel.org/oe-kbuild-all/202601211830.aBEjmEFD-lkp@intel.com/ Signed-off-by: Ted Logan <tedlogan@fb.com> Reviewed-by: David Matlack <dmatlack@google.com> Link: https://lore.kernel.org/r/20260202-vfio-selftest-only-64bit-v2-1-9c3ebb37f0f4@fb.com Signed-off-by: Alex Williamson <alex@shazbot.org> 
Only build vfio self-tests on arm64 and x86_64; these are the only
architectures where the vfio self-tests are run. Addresses compiler
warnings for format and conversions on i386.

Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202601211830.aBEjmEFD-lkp@intel.com/
Signed-off-by: Ted Logan <tedlogan@fb.com>
Reviewed-by: David Matlack <dmatlack@google.com>
Link: https://lore.kernel.org/r/20260202-vfio-selftest-only-64bit-v2-1-9c3ebb37f0f4@fb.com
Signed-off-by: Alex Williamson <alex@shazbot.org>
vfio: selftests: Drop IOMMU mapping size assertions for VFIO_TYPE1_IOMMU 2026-01-19T17:06:30+00:00 David Matlack dmatlack@google.com 2026-01-14T21:12:52+00:00 1c588bca3bd5b39c93a28a5986bf82ebfb05eec2 Drop the assertions about IOMMU mappings sizes for VFIO_TYPE1_IOMMU modes (both the VFIO mode and the iommufd compatibility mode). These assertions fail when CONFIG_IOMMUFD_VFIO_CONTAINER is enabled, since iommufd compatibility mode provides different huge page behavior than VFIO for VFIO_TYPE1_IOMMU. VFIO_TYPE1_IOMMU is an old enough interface that it's not worth changing the behavior of VFIO and iommufd to match nor care about the IOMMU mapping sizes. Cc: Jason Gunthorpe <jgg@ziepe.ca> Link: https://lore.kernel.org/kvm/20260109143830.176dc279@shazbot.org/ Signed-off-by: David Matlack <dmatlack@google.com> Link: https://lore.kernel.org/r/20260114211252.2581145-1-dmatlack@google.com Signed-off-by: Alex Williamson <alex@shazbot.org> 
Drop the assertions about IOMMU mappings sizes for VFIO_TYPE1_IOMMU
modes (both the VFIO mode and the iommufd compatibility mode). These
assertions fail when CONFIG_IOMMUFD_VFIO_CONTAINER is enabled, since
iommufd compatibility mode provides different huge page behavior than
VFIO for VFIO_TYPE1_IOMMU. VFIO_TYPE1_IOMMU is an old enough interface
that it's not worth changing the behavior of VFIO and iommufd to match
nor care about the IOMMU mapping sizes.

Cc: Jason Gunthorpe <jgg@ziepe.ca>
Link: https://lore.kernel.org/kvm/20260109143830.176dc279@shazbot.org/
Signed-off-by: David Matlack <dmatlack@google.com>
Link: https://lore.kernel.org/r/20260114211252.2581145-1-dmatlack@google.com
Signed-off-by: Alex Williamson <alex@shazbot.org>
vfio: selftests: Add vfio_dma_mapping_mmio_test 2026-01-19T17:06:30+00:00 Alex Mastro amastro@fb.com 2026-01-14T18:57:18+00:00 080723f4d4c3c6fb0720aae614deb1f30ee9ef2e Test IOMMU mapping the BAR mmaps created during vfio_pci_device_setup(). All IOMMU modes are tested: vfio_type1 variants are expected to succeed, while non-type1 modes are expected to fail. iommufd compat mode can be updated to expect success once kernel support lands. Native iommufd will not support mapping vaddrs backed by MMIO (it will support dma-buf based MMIO mapping instead). Signed-off-by: Alex Mastro <amastro@fb.com> Reviewed-by: David Matlack <dmatlack@google.com> Tested-by: David Matlack <dmatlack@google.com> Link: https://lore.kernel.org/r/20260114-map-mmio-test-v3-3-44e036d95e64@fb.com Signed-off-by: Alex Williamson <alex@shazbot.org> 
Test IOMMU mapping the BAR mmaps created during vfio_pci_device_setup().

All IOMMU modes are tested: vfio_type1 variants are expected to succeed,
while non-type1 modes are expected to fail. iommufd compat mode can be
updated to expect success once kernel support lands. Native iommufd will
not support mapping vaddrs backed by MMIO (it will support dma-buf based
MMIO mapping instead).

Signed-off-by: Alex Mastro <amastro@fb.com>
Reviewed-by: David Matlack <dmatlack@google.com>
Tested-by: David Matlack <dmatlack@google.com>
Link: https://lore.kernel.org/r/20260114-map-mmio-test-v3-3-44e036d95e64@fb.com
Signed-off-by: Alex Williamson <alex@shazbot.org>