I can confirm this on several machines in the reproducible builds zoo, running linux-image-4.9.0-8-armmp-lpae version 4.9.110-3+deb9u3.
What I've noticed is that machines running the "armmp" variant appear to be unaffected, all of the systems that failed were running the "armmp-lpae" variant. Switching to linux-image-4.9.0-8-armmp kernel variant on an affected system appears to workaround the issue, or downgrading to linux-image-4.9.0-7-armmp-lpae. One "armmp-lpae" system appears to have successfully booted, but gets many kernel messages along these lines: [ 78.638348] INFO: rcu_sched detected stalls on CPUs/tasks: [ 78.642433] 0-...: (0 ticks this GP) idle=29c/0/0 softirq=1396/1396 fqs=0 [ 78.649321] (detected by 2, t=4204 jiffies, g=283, c=282, q=41917) [ 78.655567] Task dump for CPU 0: [ 78.658770] swapper/0 R running task 0 0 0 0x00000000 [ 78.665844] rcu_sched kthread starved for 4207 jiffies! g283 c282 f0x0 RCU_GP_WAIT_FQS(3) ->state=0x1 [ 78.674972] rcu_sched S 0 7 2 0x00000000 [ 78.680506] [<c09283cc>] (__schedule) from [<c092894c>] (schedule+0x50/0xa8) [ 78.687501] [<c092894c>] (schedule) from [<c092bd50>] (schedule_timeout+0x1e0/0x464) [ 78.695231] [<c092bd50>] (schedule_timeout) from [<c02df870>] (rcu_gp_kthread+0x574/0x908) [ 78.703447] [<c02df870>] (rcu_gp_kthread) from [<c028b598>] (kthread+0x118/0x130) [ 78.710894] [<c028b598>] (kthread) from [<c0226da0>] (ret_from_fork+0x14/0x34) live well, vagrant
signature.asc
Description: PGP signature
