On Wed, Mar 04, 2026 at 08:47:15AM +0800, Yao Yuan wrote:
> On Tue, Mar 03, 2026 at 11:46:24AM +0800, Alex Mastro wrote:
> > 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 <[email protected]>
> > ---
> > The bug was missed because the test was originally run against a kernel
> > without commit afb47765f923 ("iommufd: Make vfio_compat's unmap succeed
> > if the range is already empty"). Without that fix, the unmap ioctl
> > returned -ENOENT, taking the early return before list_del_init().
> > ---
> >  tools/testing/selftests/vfio/vfio_dma_mapping_mmio_test.c | 1 -
> >  1 file changed, 1 deletion(-)
> >
> > diff --git a/tools/testing/selftests/vfio/vfio_dma_mapping_mmio_test.c 
> > b/tools/testing/selftests/vfio/vfio_dma_mapping_mmio_test.c
> > index 957a89ce7b3a..d7f25ef77671 100644
> > --- a/tools/testing/selftests/vfio/vfio_dma_mapping_mmio_test.c
> > +++ b/tools/testing/selftests/vfio/vfio_dma_mapping_mmio_test.c
> > @@ -100,7 +100,6 @@ static void do_mmio_map_test(struct iommu *iommu,
> >             iommu_unmap(iommu, &region);
> >     } else {
> 
> Hi Alex,
> 
> >             VFIO_ASSERT_NE(__iommu_map(iommu, &region), 0);
> > -           VFIO_ASSERT_NE(__iommu_unmap(iommu, &region, NULL), 0);
> 
> This is the more simply way to work w/ or w/o commit
> afb47765f923 ("iommufd: Make vfio_compat's unmap succeed if
> the range is already empty"), may worth to add this point
> into the commit message.

True, but afb47765f923 was already in the vfio tree at the time the new test
was added in 080723f4d4c3. It was just (my own) negligence that I didn't run
the selftests against <current kernel>. I wrongly assumed that running the
selftests against an older kernel that I happened to have installed on my box
was good enough.

> 
> For the changes:
> 
> Reviewed-by: Yuan Yao <[email protected]>

Thank you!

> 
> >     }
> >  }
> >
> >
> > ---
> > base-commit: 96ca4caf9066f5ebd35b561a521af588a8eb0215
> > change-id: 20260303-fix-mmio-test-d3bd688105f3
> >
> > Best regards,
> > --
> > Alex Mastro <[email protected]>
> >

Reply via email to