On 8/25/20 2:52 AM, Dmitry Shulyak wrote: > this patch fixes the issue with 0 reads. there seems to be a > regression that is not specific to uring, > regular syscall reads slowed down noticeably. Do you have a test case? Related to specific system calls, or just overall? My initial suspicion would be Yet Another one of the security fixes... -- Jens Axboe
