http://bugzilla.suse.com/show_bug.cgi?id=1131330 http://bugzilla.suse.com/show_bug.cgi?id=1131330#c5 --- Comment #5 from Michael Matz <matz@suse.com> --- Hah! I found a machine with HLE and SLE15 SP1: % lscpu | grep hle .... hle ... % rpm -q glibc glibc-2.26-15.10.x86_64 % ./pthread_mutex_destroy_ebusy pthread_mutex_destroy returned 16 (after about 20 seconds or so). So I think my theory is somewhat valid, the HLE code path is either flaky, or the program is flaky in a way more exposed in that code path, or the assumption that EBUSY must not occur is flaky. -- You are receiving this mail because: You are on the CC list for the bug.