On Thu, May 18, 2017 at 07:21:49AM +0200, Marc Burkhardt wrote:
> tonight my dmesg was flooded with mesages like
>
> vmscan: shrink_slab: ext4_es_scan+0x0/0x150 negative objects to delete
> nr=-2147483624
>
> Is that an integer overflow happening in ext4?
>
> It's the first time I see this message. Any help on how to debug/reprocude
> this
> are appreciated. Please advice if you want me to investigate this.
I haven't attempted to debug nor reproduce it, but what I can tell you
is that it does not not have anything to with ext4. I've seen similar
messages with a completely different slab, on 4.9.26:
[367594.725081] vmscan: shrink_slab: super_cache_scan+0x0/0x19c negative
objects to delete nr=-2147482285
[367595.046073] vmscan: shrink_slab: super_cache_scan+0x0/0x19c negative
objects to delete nr=-2147479427
[367595.279228] vmscan: shrink_slab: super_cache_scan+0x0/0x19c negative
objects to delete nr=-2147482317
[367595.459529] vmscan: shrink_slab: super_cache_scan+0x0/0x19c negative
objects to delete nr=-2147482353
[367595.497191] vmscan: shrink_slab: super_cache_scan+0x0/0x19c negative
objects to delete nr=-2147482386
[367595.521578] vmscan: shrink_slab: super_cache_scan+0x0/0x19c negative
objects to delete nr=-2147482413
[367595.551109] vmscan: shrink_slab: super_cache_scan+0x0/0x19c negative
objects to delete nr=-2147482501
[367598.344400] vmscan: shrink_slab: super_cache_scan+0x0/0x19c negative
objects to delete nr=-2147482458
[367598.369103] vmscan: shrink_slab: super_cache_scan+0x0/0x19c negative
objects to delete nr=-2147482493
[367598.403148] vmscan: shrink_slab: super_cache_scan+0x0/0x19c negative
objects to delete nr=-2147482521
[367598.422815] vmscan: shrink_slab: super_cache_scan+0x0/0x19c negative
objects to delete nr=-2147482611
[367598.524128] vmscan: shrink_slab: super_cache_scan+0x0/0x19c negative
objects to delete nr=-2147483238
[367601.554775] vmscan: shrink_slab: super_cache_scan+0x0/0x19c negative
objects to delete nr=-2147482245
[367601.582922] vmscan: shrink_slab: super_cache_scan+0x0/0x19c negative
objects to delete nr=-2147482279
[367601.620175] vmscan: shrink_slab: super_cache_scan+0x0/0x19c negative
objects to delete nr=-2147482307
[367602.958946] vmscan: shrink_slab: super_cache_scan+0x0/0x19c negative
objects to delete nr=-2147479516
[367603.630417] vmscan: shrink_slab: super_cache_scan+0x0/0x19c negative
objects to delete nr=-2147482412
[367603.746885] vmscan: shrink_slab: super_cache_scan+0x0/0x19c negative
objects to delete nr=-2147482512
[367603.769490] vmscan: shrink_slab: super_cache_scan+0x0/0x19c negative
objects to delete nr=-2147482217
[367604.155461] vmscan: shrink_slab: super_cache_scan+0x0/0x19c negative
objects to delete nr=-2147479940
[367604.174624] vmscan: shrink_slab: super_cache_scan+0x0/0x19c negative
objects to delete nr=-2147482635
[367604.197573] vmscan: shrink_slab: super_cache_scan+0x0/0x19c negative
objects to delete nr=-2147482595
I don't see any fixes/changes to mm/vmscan.c in newer 4.9-stable kernels
other than these ones which were already merged v4.9.14:
$ git shortlog v4.9..v4.9.30 -- mm/vmscan.c
Michal Hocko (3):
mm, memcg: fix the active list aging for lowmem requests when memcg is
enabled
mm, vmscan: cleanup lru size claculations
mm, vmscan: consider eligible zones in get_scan_count
Shaohua Li (1):
mm/vmscan.c: set correct defer count for shrinker
Perhaps one of the above people or someone else in linux-mm recognizes this.
