perf: perf_fuzzer lockup on 4.12-rc1

From: Vince Weaver
Date: Fri May 19 2017 - 16:43:03 EST


Hello

the lack of perf_fuzzer reports hasn't been due to lack of issues, just my
fuzzing machines were needed by students. The semester is over so back to
fuzzing.

My haswell machine still can't go more than a day or so without locking
hard. Non-reproducible as always.

Here's a hang on 4.12-rc1. The backtrace takes an interesting path
through bpf code, I don't think I've seen this particular one before.

Vince


[57549.283910] INFO: rcu_sched detected stalls on CPUs/tasks:
[57549.289788] 5-...: (0 ticks this GP) idle=71e/140000000000000/0 softirq=4813112/4813112 fqs=2626
[57549.299316] (detected by 6, t=5255 jiffies, g=5128379, c=5128378, q=114)
[57549.306599] Sending NMI from CPU 6 to CPUs 5:
[57549.306687] NMI backtrace for cpu 5
[57549.306688] CPU: 5 PID: 20120 Comm: perf_fuzzer Tainted: G W 4.12.0-rc1+ #27
[57549.306688] Hardware name: LENOVO 10AM000AUS/SHARKBAY, BIOS FBKT72AUS 01/26/2014
[57549.306688] task: ffff880037036000 task.stack: ffffc90008a94000
[57549.306689] RIP: 0033:0x7fa028f6a030
[57549.306689] RSP: 002b:00007ffc93d544b8 EFLAGS: 00000206
[57549.306689] RAX: 0000000000000027 RBX: 0000000000000027 RCX: 00007fa0292210a8
[57549.306690] RDX: 0000000000000027 RSI: 00007ffc93d5449c RDI: 00007fa029221600
[57549.306690] RBP: 00007ffc93d544f0 R08: 00007fa0292210a4 R09: 00007fa029221120
[57549.306690] R10: 00000000ffffffff R11: 0000000000000246 R12: 000055f5c791bc10
[57549.306691] R13: 00007ffc93d56970 R14: 0000000000000000 R15: 0000000000000000
[57549.306691] FS: 00007fa02942f700(0000) GS:ffff88011eb40000(0000) knlGS:0000000000000000
[57549.306691] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[57549.306691] CR2: 000055f5cf90f058 CR3: 00000001177e5000 CR4: 00000000001407e0
[57549.306692] DR0: 0000001000000000 DR1: 0000000000000000 DR2: 0000000000000000
[57549.306692] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000600
[58184.565805] NMI watchdog: BUG: soft lockup - CPU#0 stuck for 22s!
[(journald):21894]
[58184.646703] CPU: 0 PID: 21894 Comm: (journald) Tainted: G W 4.12.0-rc1+ #27
[58184.655370] Hardware name: LENOVO 10AM000AUS/SHARKBAY, BIOS FBKT72AUS 01/26/2014
[58184.663264] task: ffff8800cf5e0180 task.stack: ffffc9000b51c000
[58184.669582] RIP: 0010:smp_call_function_many+0x1d2/0x230
[58184.675238] RSP: 0018:ffffc9000b51fc60 EFLAGS: 00000202 ORIG_RAX: ffffffffffffff10
[58184.683316] RAX: 0000000000000003 RBX: 0000000000000200 RCX: 0000000000000002
[58184.690960] RDX: ffff88011ea9ee38 RSI: 0000000000000000 RDI: ffff88011ea1bcc8
[58184.698627] RBP: ffffc9000b51fc98 R08: fffffffffffffffc R09: 00000000000000fe
[58184.706247] R10: 0000000000000008 R11: 0000000000000002 R12: ffff88011ea1bcc8
[58184.713926] R13: ffff88011ea1bcc0 R14: ffffffff81069a30 R15: 0000000000000000
[58184.721570] FS: 00007fb170bff940(0000) GS:ffff88011ea00000(0000) knlGS:00000
[58184.721570] FS: 00007fb170bff940(0000) GS:ffff88011ea00000(0000) knlGS:0000000000000000
[58184.730202] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[58184.736372] CR2: 00005626954ee000 CR3: 00000000ce653000 CR4: 00000000001406f0
[58184.743990] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[58184.751589] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000600
[58184.759191] Call Trace:
[58184.761825] ? leave_mm+0xd0/0xd0
[58184.765358] on_each_cpu+0x2d/0x60
[58184.769002] flush_tlb_kernel_range+0x3c/0x60
[58184.773652] __purge_vmap_area_lazy+0x50/0xc0
[58184.778305] vm_unmap_aliases+0x101/0x140
[58184.782573] change_page_attr_set_clr+0xd1/0x400
[58184.787500] set_memory_ro+0x29/0x30
[58184.791323] bpf_prog_select_runtime+0x2e/0xc0
[58184.796083] bpf_prepare_filter+0x366/0x3d0
[58184.800563] bpf_prog_create_from_user+0xbf/0x120
[58184.805588] ? watchdog_nmi_disable+0x70/0x70
[58184.810251] do_seccomp+0xf7/0x620
[58184.813901] SyS_seccomp+0xe/0x10
[58184.817461] entry_SYSCALL_64_fastpath+0x1e/0xa9