On 2/23/21 6:36 AM, Bruno Goncalves wrote: > Hello, > > In this run a lot of tests failed due to a known issue with SELinux policy, but 1 failure reported on LTP io_uring01 test when running on ppc64le might be a kernel bug [1]. > Note the sefgault (io_uring01[138555]: User access of kernel address (c0000000005a0258) - exploit attempt? (uid: 0)) on console log [2] when executing the test. The test logs are [3]. > > [1] https://arr-cki-prod-datawarehouse-public.s3.amazonaws.com/datawarehouse-public/2021/02/23/624525/build_ppc64le_redhat%3A1110530/tests/LTP/9600734_ppc64le_1_syscalls.fail.log > [2] https://arr-cki-prod-datawarehouse-public.s3.amazonaws.com/datawarehouse-public/2021/02/23/624525/build_ppc64le_redhat%3A1110530/tests/9600734_ppc64le_1_console.log > [3] https://arr-cki-prod-datawarehouse-public.s3.amazonaws.com/index.html?prefix=datawarehouse-public/2021/02/23/624525/build_ppc64le_redhat%3A1110530/tests/LTP/ That's my fault, for some reason I totally overlooked parisc and powerpc in the PF_IO_WORKER change. Current for-next should work for you, this is the change that fixes it: https://git.kernel.dk/cgit/linux-block/commit/?h=io_uring-worker.v3&id=e4595c30ec3053a15d12615195b7a8726f9bee79 -- Jens Axboe
