Re: kernel BUG at mm/huge_memory.c:1187!

[Kirill A. Shutemov]

On Sun, Oct 09, 2016 at 03:24:27PM +0800, Wenwei Tao wrote:
> Hi,
> 
> I open the Transparent  huge page and run the system and hit the bug
> in huge_memory.c:
> 
> static void __split_huge_page_refcount(struct page *page)
>                               .
>                               .
>                               .
> 
>      /* tail_page->_mapcount cannot change */
>      BUG_ON(page_mapcount(page_tail) < 0);
>                                .
>                                .
> 
> In my understanding, the THP's tail page's mapcount is initialized to
> -1,  page_mapcout(page_tail) should be 0.
> Did anyone meet the same issue?
> 
> Thanks.
> 
> 2016-09-28 02:12:08 [810422.485203] ------------[ cut here ]------------
> 2016-09-28 02:12:08 [810422.489974] kernel BUG at mm/huge_memory.c:1187!
> 2016-09-28 02:12:08 [810422.494742] invalid opcode: 0000 [#1] SMP
> 2016-09-28 02:12:08 [810422.499034] last sysfs file:
> /sys/devices/system/cpu/online
> 2016-09-28 02:12:08 [810422.504757] CPU 31
> 2016-09-28 02:12:08 [810422.506775] Modules linked in: 8021q garp
> bridge stp llc dell_rbu ipmi_devintf ipmi_si ipmi_msghandler bonding
> ipv6 microcode  dca power_meter ext4 mbcache jbd2 ahci wmi dm_mirror
> dm_region_hash dm_log dm_mod
> 2016-09-28 02:12:08 [810422.571439]
> 2016-09-28 02:12:08 [810422.573088] Pid: 10729, comm: observer
> Tainted: G        W  ----------------   2.6.32-220.23.2.el6.x86_64

The kernel is already tianted with warning.

And that is way to old kernel -- from RHEL 6.2, almost 5 years old.
Contact your OS vendor.

-- 
 Kirill A. Shutemov