tensorflow is vulnerable to denial of service (DoS) attacks. An attacker is able to cause denial of service conditions by providing empty inputs through SparseFillEmptyRowsGrad
.
github.com/advisories/GHSA-hq7g-wwwp-q46h
github.com/tensorflow/tensorflow/blob/master/tensorflow/core/kernels/sparse_fill_empty_rows_op_gpu.cu.cc
github.com/tensorflow/tensorflow/commit/507941f3dd0639d7a9c51ba47c8459f3e0e0dc85
github.com/tensorflow/tensorflow/commit/8c3f7009710749e0d5a86bbc394fbf405b2e9f5c
github.com/tensorflow/tensorflow/commit/a838d9e739b4a4b948ae867b28aa17a5dcb45c38
github.com/tensorflow/tensorflow/commit/af4a6a3c8b95022c351edae94560acc61253a1b8
github.com/tensorflow/tensorflow/pull/58559
github.com/tensorflow/tensorflow/pull/58562
github.com/tensorflow/tensorflow/pull/58565
github.com/tensorflow/tensorflow/security/advisories/GHSA-hq7g-wwwp-q46h