[PATCH] iommufd/selftest: Rework TEST_LENGTH to test min_size explicitly

[PATCH] iommufd/selftest: Rework TEST_LENGTH to test min_size explicitly

[Nicolin Chen]

TEST_LENGTH passing ".size = sizeof(struct _struct) - 1" expects -EINVAL
from "if (ucmd.user_size < op->min_size)" check in iommufd_fops_ioctl().
This has been working when min_size is exactly the size of the structure.

However, if the size of the structure becomes larger than min_size, i.e.
the passing size above is larger than min_size, that min_size sanity no
longer works.

Since the first test in TEST_LENGTH() was to test that min_size sanity
routine, rework it to support a min_size calculation, rather than using
the full size of the structure.

Signed-off-by: Nicolin Chen <nicolinc@nvidia.com>
---
Hi Jason/Kevin,

This was a part of the nesting series. Its link in v4:
https://lore.kernel.org/linux-iommu/20230921075138.124099-13-yi.l.liu@intel.com/

I just realized that this should go in prior to the nesting series.
One of the nesting patches changes the IOMMU_HWPT_ALLOC structure,
which would break the cmd_length test without this patch.

Thanks!
Nicolin

 tools/testing/selftests/iommu/iommufd.c | 29 ++++++++++++++-----------
 1 file changed, 16 insertions(+), 13 deletions(-)

diff --git a/tools/testing/selftests/iommu/iommufd.c b/tools/testing/selftests/iommu/iommufd.c
index c5eca2fee42c..6323153d277b 100644
--- a/tools/testing/selftests/iommu/iommufd.c
+++ b/tools/testing/selftests/iommu/iommufd.c
@@ -86,12 +86,13 @@ TEST_F(iommufd, cmd_fail)
 
 TEST_F(iommufd, cmd_length)
 {
-#define TEST_LENGTH(_struct, _ioctl)                                     \
+#define TEST_LENGTH(_struct, _ioctl, _last)                              \
 	{                                                                \
+		size_t min_size = offsetofend(struct _struct, _last);    \
 		struct {                                                 \
 			struct _struct cmd;                              \
 			uint8_t extra;                                   \
-		} cmd = { .cmd = { .size = sizeof(struct _struct) - 1 }, \
+		} cmd = { .cmd = { .size = min_size - 1 },               \
 			  .extra = UINT8_MAX };                          \
 		int old_errno;                                           \
 		int rc;                                                  \
@@ -112,17 +113,19 @@ TEST_F(iommufd, cmd_length)
 		}                                                        \
 	}
 
-	TEST_LENGTH(iommu_destroy, IOMMU_DESTROY);
-	TEST_LENGTH(iommu_hw_info, IOMMU_GET_HW_INFO);
-	TEST_LENGTH(iommu_hwpt_alloc, IOMMU_HWPT_ALLOC);
-	TEST_LENGTH(iommu_ioas_alloc, IOMMU_IOAS_ALLOC);
-	TEST_LENGTH(iommu_ioas_iova_ranges, IOMMU_IOAS_IOVA_RANGES);
-	TEST_LENGTH(iommu_ioas_allow_iovas, IOMMU_IOAS_ALLOW_IOVAS);
-	TEST_LENGTH(iommu_ioas_map, IOMMU_IOAS_MAP);
-	TEST_LENGTH(iommu_ioas_copy, IOMMU_IOAS_COPY);
-	TEST_LENGTH(iommu_ioas_unmap, IOMMU_IOAS_UNMAP);
-	TEST_LENGTH(iommu_option, IOMMU_OPTION);
-	TEST_LENGTH(iommu_vfio_ioas, IOMMU_VFIO_IOAS);
+	TEST_LENGTH(iommu_destroy, IOMMU_DESTROY, id);
+	TEST_LENGTH(iommu_hw_info, IOMMU_GET_HW_INFO, __reserved);
+	TEST_LENGTH(iommu_hwpt_alloc, IOMMU_HWPT_ALLOC, __reserved);
+	TEST_LENGTH(iommu_ioas_alloc, IOMMU_IOAS_ALLOC, out_ioas_id);
+	TEST_LENGTH(iommu_ioas_iova_ranges, IOMMU_IOAS_IOVA_RANGES,
+		    out_iova_alignment);
+	TEST_LENGTH(iommu_ioas_allow_iovas, IOMMU_IOAS_ALLOW_IOVAS,
+		    allowed_iovas);
+	TEST_LENGTH(iommu_ioas_map, IOMMU_IOAS_MAP, iova);
+	TEST_LENGTH(iommu_ioas_copy, IOMMU_IOAS_COPY, src_iova);
+	TEST_LENGTH(iommu_ioas_unmap, IOMMU_IOAS_UNMAP, length);
+	TEST_LENGTH(iommu_option, IOMMU_OPTION, val64);
+	TEST_LENGTH(iommu_vfio_ioas, IOMMU_VFIO_IOAS, __reserved);
 #undef TEST_LENGTH
 }
 
-- 
2.42.0

RE: [PATCH] iommufd/selftest: Rework TEST_LENGTH to test min_size explicitly

[Tian, Kevin]

> From: Nicolin Chen <nicolinc@nvidia.com>
> Sent: Sunday, October 15, 2023 3:47 PM
> 
> TEST_LENGTH passing ".size = sizeof(struct _struct) - 1" expects -EINVAL
> from "if (ucmd.user_size < op->min_size)" check in iommufd_fops_ioctl().
> This has been working when min_size is exactly the size of the structure.
> 
> However, if the size of the structure becomes larger than min_size, i.e.
> the passing size above is larger than min_size, that min_size sanity no
> longer works.
> 
> Since the first test in TEST_LENGTH() was to test that min_size sanity
> routine, rework it to support a min_size calculation, rather than using
> the full size of the structure.
> 
> Signed-off-by: Nicolin Chen <nicolinc@nvidia.com>

Reviewed-by: Kevin Tian <kevin.tian@intel.com>

Re: [PATCH] iommufd/selftest: Rework TEST_LENGTH to test min_size explicitly

[Jason Gunthorpe]

On Sun, Oct 15, 2023 at 12:46:48AM -0700, Nicolin Chen wrote:
> TEST_LENGTH passing ".size = sizeof(struct _struct) - 1" expects -EINVAL
> from "if (ucmd.user_size < op->min_size)" check in iommufd_fops_ioctl().
> This has been working when min_size is exactly the size of the structure.
> 
> However, if the size of the structure becomes larger than min_size, i.e.
> the passing size above is larger than min_size, that min_size sanity no
> longer works.
> 
> Since the first test in TEST_LENGTH() was to test that min_size sanity
> routine, rework it to support a min_size calculation, rather than using
> the full size of the structure.
> 
> Signed-off-by: Nicolin Chen <nicolinc@nvidia.com>
> ---
> Hi Jason/Kevin,
> 
> This was a part of the nesting series. Its link in v4:
> https://lore.kernel.org/linux-iommu/20230921075138.124099-13-yi.l.liu@intel.com/
> 
> I just realized that this should go in prior to the nesting series.
> One of the nesting patches changes the IOMMU_HWPT_ALLOC structure,
> which would break the cmd_length test without this patch.
> 
> Thanks!
> Nicolin
> 
>  tools/testing/selftests/iommu/iommufd.c | 29 ++++++++++++++-----------
>  1 file changed, 16 insertions(+), 13 deletions(-)

Applied to iommufd for-next

Thanks,
Jason