On Mon, Apr 30, 2018 at 08:30:43AM -0700, Dave Hansen wrote: > On 04/28/2018 12:05 AM, Ingo Molnar wrote: > > In the above kernel that was missing the PROT_EXEC fix I was repeatedly running > > the 64-bit and 32-bit testcases as non-root and as root as well, until I got a > > hang in the middle of a 32-bit test running as root: > > > > test 7 PASSED (iteration 19) > > test 8 PASSED (iteration 19) > > test 9 PASSED (iteration 19) > > > > < test just hangs here > > > For the hang, there is a known issue with the use of printf() in the > signal handler and a resulting deadlock. I *thought* there was a patch > merged to fix this from Ram Pai or one of the other IBM folks. Yes. there is a patch. unfortunately that patch assumes the selftest has been moved into selftests/vm directory. One option is -- I merge your changes in my selftest patchset, and send the entire series for upstream merge. Or you can manually massage-in the specific fix. The patch is "selftests/vm: Fix deadlock in protection_keys.c" https://patchwork.ozlabs.org/patch/864394/ Let me know, -- Ram Pai
