On Sun, Jul 28, 2024 at 7:47 AM Drew Parsons <dpars...@debian.org> wrote: > > On 2024-07-25 12:35, Drew Parsons wrote: > > Given the timing of the test failure, my guess is that uid-wrapper is a > > casuality of the time_64 transition. It looks like something is amiss > > with the __NR_gettimeofday system calls on 32-bit ARM. > > > > cc:ing ARM porters. Can you advise on the 32-bit arm test failure > > reported in Bug#1069425 ? > > SYS_gettimeofday is returning zero time. > > I've uploaded an NMU for uid-wrapper 1.3.0-5.1 to skip the syscall > gettimeofday test on __arm__, since armel and armhf are returning > tv_sec=0 from the syscall. Obviously replace the patch with a better > fix if one is identified later.
I believe that will disable the test on aarch64, too. Is that intended? If not, then I believe you need to pivot on __arm64__, __aarch32__ and __aarch64__, too. Jeff
