On 10/2/21 2:44 AM, Pavel Begunkov wrote: > For some reason non-off IORING_OP_TIMEOUT always fails links, it's > pretty inconvenient and unnecessary limits chaining after it to hard > linking, which is far from ideal, e.g. doesn't pair well with timeout > cancellation. Prevent it and treat -ETIME as success. That seems like a sane addition, but I'm not a huge fan of the #define IORING_TIMEOUT_DONT_FAIL (1U << 5) name, as it isn't very descriptive. Don't fail what? Maybe #define IORING_TIMEOUT_ETIME_SUCCESS instead? At least that tells the story of -ETIME being considered success, hence not breaking a link. -- Jens Axboe
