On Wed, 2014-05-28 at 17:33 -0700, Hugh Dickins wrote:
> On Wed, 28 May 2014, Michael Ellerman wrote:
> > Linux Blade312-5 3.15.0-rc7 #306 SMP Wed May 28 17:51:18 EST 2014 ppc64
> >
> > [watchdog] 27853 iterations. [F:22642 S:5174 HI:1276]
> > ------------[ cut here ]------------
> > kernel BUG at /home/michael/mmk-build/flow/mm/memory.c:1489!
> > cpu 0xc: Vector: 700 (Program Check) at [c000000384eaf960]
> > pc: c0000000001ad6f0: .follow_page_mask+0x90/0x650
> > lr: c0000000001ad6d8: .follow_page_mask+0x78/0x650
> > sp: c000000384eafbe0
> > msr: 8000000000029032
> > current = 0xc0000003c27e1bc0
> > paca = 0xc000000001dc3000 softe: 0 irq_happened: 0x01
> > pid = 20800, comm = trinity-c12
> > kernel BUG at /home/michael/mmk-build/flow/mm/memory.c:1489!
> > enter ? for help
> > [c000000384eafcc0] c0000000001e5514 .SyS_move_pages+0x524/0x7d0
> > [c000000384eafe30] c00000000000a1d8 syscall_exit+0x0/0x98
> > --- Exception: c01 (System Call) at 00003fff795f30a8
> > SP (3ffff958f290) is in userspace
> >
> > I've left it in the debugger, can dig into it a bit more tomorrow
> > if anyone has any clues.
>
> Thanks for leaving it overnight, but this one is quite obvious,
> so go ahead and reboot whenever suits you.
>
> Trinity didn't even need to do anything bizarre to get this: that
> ordinary path simply didn't get tried on powerpc or ia64 before.
>
> Here's a patch which should fix it for you, but I believe leaves
> a race in common with other architectures. I must turn away to
> other things, and hope Naoya-san can fix up the locking separately
> (or point out why it's already safe).
>
> [PATCH] mm: fix move_pages follow_page huge_addr BUG
>
> v3.12's e632a938d914 ("mm: migrate: add hugepage migration code to
> move_pages()") is okay on most arches, but on follow_huge_addr-style
> arches ia64 and powerpc, it hits my old BUG_ON(flags & FOLL_GET)
> from v2.6.15 deceb6cd17e6 ("mm: follow_page with inner ptlock").
>
> The point of the BUG_ON was that nothing needed FOLL_GET there at
> the time, and it was not clear that we have sufficient locking to
> use get_page() safely here on the outside - maybe the page found has
> already been freed and even reused when follow_huge_addr() returns.
>
> I suspect that e632a938d914's use of get_page() after return from
> follow_huge_pmd() has the same problem: what prevents a racing
> instance of move_pages() from already migrating away and freeing
> that page by then? A reference to the page should be taken while
> holding suitable lock (huge_pte_lockptr?), to serialize against
> concurrent migration.
>
> But I'm not prepared to rework the hugetlb locking here myself;
> so for now just supply a patch to copy e632a938d914's get_page()
> after follow_huge_pmd() to after follow_huge_addr(): removing
> the BUG_ON(flags & FOLL_GET), but probably leaving a race.
Thanks for the detailed explanation Hugh.
Unfortunately I don't know our mm/hugetlb code well enough to give you a good
answer. Ben had a quick look at our follow_huge_addr() and thought it looked
"fishy". He suggested something like what we do in gup_pte_range() with
page_cache_get_speculative() might be in order.
Applying your patch and running trinity pretty immediately results in the
following, which looks related (sys_move_pages() again) ?
Unable to handle kernel paging request for data at address 0xf2000f80000000
Faulting instruction address: 0xc0000000001e29bc
cpu 0x1b: Vector: 300 (Data Access) at [c0000003c70f76f0]
pc: c0000000001e29bc: .remove_migration_pte+0x9c/0x320
lr: c0000000001e29b8: .remove_migration_pte+0x98/0x320
sp: c0000003c70f7970
msr: 8000000000009032
dar: f2000f80000000
dsisr: 40000000
current = 0xc0000003f9045800
paca = 0xc000000001dc6c00 softe: 0 irq_happened: 0x01
pid = 3585, comm = trinity-c27
enter ? for help
[c0000003c70f7a20] c0000000001bce88 .rmap_walk+0x328/0x470
[c0000003c70f7ae0] c0000000001e2904 .remove_migration_ptes+0x44/0x60
[c0000003c70f7b80] c0000000001e4ce8 .migrate_pages+0x6d8/0xa00
[c0000003c70f7cc0] c0000000001e55ec .SyS_move_pages+0x5dc/0x7d0
[c0000003c70f7e30] c00000000000a1d8 syscall_exit+0x0/0x98
--- Exception: c01 (System Call) at 00003fff7b2b30a8
SP (3fffe09728a0) is in userspace
1b:mon>
I've hit it twice in two runs:
Unable to handle kernel paging request for data at address 0xf2400f00000000
Faulting instruction address: 0xc0000000001e2a3c
cpu 0xd: Vector: 300 (Data Access) at [c00000038a4bf6f0]
pc: c0000000001e2a3c: .remove_migration_pte+0x9c/0x320
lr: c0000000001e2a38: .remove_migration_pte+0x98/0x320
sp: c00000038a4bf970
msr: 8000000000009032
dar: f2400f00000000
dsisr: 40000000
current = 0xc0000003acd9e680
paca = 0xc000000001dc3400 softe: 0 irq_happened: 0x01
pid = 13334, comm = trinity-c13
enter ? for help
[c00000038a4bfa20] c0000000001bcf08 .rmap_walk+0x328/0x470
[c00000038a4bfae0] c0000000001e2984 .remove_migration_ptes+0x44/0x60
[c00000038a4bfb80] c0000000001e4d68 .migrate_pages+0x6d8/0xa00
[c00000038a4bfcc0] c0000000001e566c .SyS_move_pages+0x5dc/0x7d0
[c00000038a4bfe30] c00000000000a1d8 syscall_exit+0x0/0x98
--- Exception: c01 (System Call) at 00003fff79df30a8
SP (3fffda95d500) is in userspace
d:mon>
If I tell trinity to skip sys_move_pages() it runs for hours.
cheers
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxx. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>