Fix for integer overflow in contiguous-split

[jbrennan333]

This is a fix for an integer overflow in contiguous_split.cu::copy_data.

We encountered this problem when running a large query with the spark-rapids plug-in. The plug-in OOM callback was reporting negative numbers for some failed allocations. It treats the allocation size as a long. The allocations were coming from contiguous_split with a large number of rows (399,000,000) of type UINT64.

With help from @nvdbaranec, I was able to isolate the problem to this line:

size_t const bytes = buf.num_elements * buf.element_size;

The two operands are both INT32. So the result of the multiply is being sign-extended when assigning to bytes.

The fix is to cast these to size_t before the multiply.

[codecov]

Codecov Report

Merging #10437 (a5acc9f) into branch-22.04 (4596244) will increase coverage by 0.04%.
The diff coverage is 100.00%.

@@               Coverage Diff                @@
##           branch-22.04   #10437      +/-   ##
================================================
+ Coverage         86.13%   86.18%   +0.04%     
================================================
  Files               139      139              
  Lines             22438    22468      +30     
================================================
+ Hits              19328    19363      +35     
+ Misses             3110     3105       -5     

Impacted Files	Coverage Δ
python/cudf/cudf/core/tools/numeric.py	89.24% <100.00%> (+0.11%)	⬆️
python/dask_cudf/dask_cudf/backends.py	86.44% <100.00%> (+1.47%)	⬆️
...ython/dask_cudf/dask_cudf/io/tests/test_parquet.py	100.00% <100.00%> (ø)
python/cudf/cudf/core/column/string.py	88.39% <0.00%> (+0.12%)	⬆️
python/cudf/cudf/core/groupby/groupby.py	91.57% <0.00%> (+0.22%)	⬆️
python/cudf/cudf/core/column/numerical.py	95.28% <0.00%> (+0.29%)	⬆️
python/cudf/cudf/core/tools/datetimes.py	84.49% <0.00%> (+0.30%)	⬆️
python/cudf/cudf/core/column/lists.py	90.56% <0.00%> (+0.47%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update d4ce5d5...a5acc9f. Read the comment docs.

[PointKernel]

Looks great. Thanks for helping clean up the code!

[PointKernel]

rerun tests

[jbrennan333]

They python build failure looks like build system issue:

[](https://gpuci.gpuopenanalytics.com/job/rapidsai/job/gpuci/job/cudf/job/prb/job/cudf-cpu-python-build/7219/console#L223)[](https://gpuci.gpuopenanalytics.com/job/rapidsai/job/gpuci/job/cudf/job/prb/job/cudf-cpu-python-build/7219/console#L224)
15:23:50 ‘EC2 (aws-b) - runner-m5d2xl (i-0fcffb338c35da323)’ doesn’t have label ‘cpu4’
15:23:50 ‘Jenkins’ doesn’t have label ‘cpu4’
15:49:46 [rapidsai » gpuci » cudf » prb » cudf-cpu-python-build » 11.5,3.9](https://gpuci.gpuopenanalytics.com/job/rapidsai/job/gpuci/job/cudf/job/prb/job/cudf-cpu-python-build/CUDA=11.5,PYTHON=3.9/)[](https://gpuci.gpuopenanalytics.com/job/rapidsai/job/gpuci/job/cudf/job/prb/job/cudf-cpu-python-build/7219/console#L225) completed with result FAILURE
15:49:46 [rapidsai » gpuci » cudf » prb » cudf-cpu-python-build » 11.5,3.8](https://gpuci.gpuopenanalytics.com/job/rapidsai/job/gpuci/job/cudf/job/prb/job/cudf-cpu-python-build/CUDA=11.5,PYTHON=3.8/) completed with result FAILURE

[jbrennan333]

Thanks for the review @ttnghia!

The build failures appear to be unrelated - do we need to wait until they are resolved before we can merge this?

[PointKernel]

The CI failure fix has been merged (see #10442). @jbrennan333 Could you merge with the latest branch-22.04?

[PointKernel]

@gpucibot merge

[jbrennan333]

Thanks for the reviews @PointKernel, @nvdbaranec, and @ttnghia! And thanks for testing the patch @nvdbaranec!