* Re: [lkp] [mm] 7d2eba0557: BUG: unable to handle kernel NULL pointer dereference at (null)
[not found] <8737tue3mn.fsf@yhuang-dev.intel.com>
@ 2016-01-19 8:02 ` Cyrill Gorcunov
2016-01-19 22:30 ` David Rientjes
0 siblings, 1 reply; 5+ messages in thread
From: Cyrill Gorcunov @ 2016-01-19 8:02 UTC (permalink / raw)
To: kernel test robot
Cc: Ebru Akagunduz, lkp, LKML, Andrew Morton, Dan Carpenter,
Michal Hocko, Johannes Weiner, Hugh Dickins, Aneesh Kumar K.V,
Vlastimil Babka, David Rientjes, Mel Gorman, Xie XiuQi,
Joonsoo Kim, Andrea Arcangeli, Naoya Horiguchi, Rik van Riel,
Kirill A. Shutemov, Linus Torvalds
On Tue, Jan 19, 2016 at 09:53:36AM +0800, kernel test robot wrote:
> FYI, we noticed the below changes on
>
> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
> commit 7d2eba0557c18f7522b98befed98799990dd4fdb ("mm: add tracepoint for scanning pages")
>
>
> +-----------------------------------------------------------+------------+------------+
> | | cb5490a5ee | 7d2eba0557 |
> +-----------------------------------------------------------+------------+------------+
> | boot_successes | 0 | 0 |
> | boot_failures | 12 | 12 |
> | BUG:kernel_test_oversize | 10 | |
> | Kernel_panic-not_syncing:Attempted_to_kill_init!exitcode= | 2 | 4 |
> | BUG:unable_to_handle_kernel | 0 | 8 |
> | Oops | 0 | 8 |
> | RIP:khugepaged_scan_pmd | 0 | 8 |
> | Kernel_panic-not_syncing:Fatal_exception | 0 | 8 |
> | backtrace:khugepaged | 0 | 8 |
> +-----------------------------------------------------------+------------+------------+
>
>
>
> [ 28.528131] systemd-journald[162]: Vacuuming...
> [ 28.543867] systemd-journald[162]: Vacuuming done, freed 0 bytes
> [ 28.548220] systemd-journald[162]: Flushing /dev/kmsg...
> [ 28.552721] BUG: unable to handle kernel NULL pointer dereference at (null)
> [ 28.553471] IP: [<ffffffffadeed16b>] khugepaged_scan_pmd+0x33b/0xaa0
> [ 28.553480] PGD 0
> [ 28.553484] Oops: 0000 [#1] PREEMPT SMP
> [ 28.553490] Modules linked in: autofs4
> [ 28.553497] CPU: 1 PID: 25 Comm: khugepaged Not tainted 4.4.0-05704-g7d2eba0 #1
> [ 28.553498] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Debian-1.8.2-1 04/01/2014
> [ 28.553501] task: ffff880035b75000 ti: ffff880035b78000 task.ti: ffff880035b78000
> [ 28.553503] RIP: 0010:[<ffffffffadeed16b>] [<ffffffffadeed16b>] khugepaged_scan_pmd+0x33b/0xaa0
> [ 28.553508] RSP: 0018:ffff880035b7bd20 EFLAGS: 00010246
> [ 28.553510] RAX: 0000000000000000 RBX: 0000000000000002 RCX: ffff880035b7bdf0
> [ 28.553512] RDX: ffff880000000000 RSI: 0000000000000000 RDI: 800000002f8000e7
> [ 28.553514] RBP: ffff880035b7bda0 R08: ffff8800155a6448 R09: 0000000000000000
> [ 28.553516] R10: 0000000000000371 R11: ffff8800354d1120 R12: 0000000000000000
> [ 28.553517] R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000001
> [ 28.553520] FS: 0000000000000000(0000) GS:ffff880036800000(0000) knlGS:0000000000000000
> [ 28.553522] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [ 28.553524] CR2: 0000000000000000 CR3: 00000000157a8000 CR4: 00000000000406a0
> [ 28.553528] Stack:
> [ 28.553530] ffff880035b75bd8 ffff880035b7bdf0 ffffffffadd3a94f ffff880035b75000
> [ 28.553535] ffff880015751510 0000000000000b54 ffff880035b7be10 ffff880035b7bd78
> [ 28.553541] 0000000000000000 00007fd735600000 ffff8800354d1000 0000000000000001
> [ 28.553546] Call Trace:
> [ 28.553552] [<ffffffffadd3a94f>] ? __lock_is_held+0x6f/0xa0
> [ 28.553556] [<ffffffffadeee2ea>] khugepaged+0xa1a/0xac0
> [ 28.553560] [<ffffffffadd30bd0>] ? woken_wake_function+0x30/0x30
> [ 28.553563] [<ffffffffadeed8d0>] ? khugepaged_scan_pmd+0xaa0/0xaa0
> [ 28.553568] [<ffffffffadcf9bb2>] kthread+0x132/0x160
> [ 28.553574] [<ffffffffadcf9a80>] ? kthread_unpark+0x80/0x80
> [ 28.553578] [<ffffffffae764c9f>] ret_from_fork+0x3f/0x70
> [ 28.553581] [<ffffffffadcf9a80>] ? kthread_unpark+0x80/0x80
> [ 28.553583] Code: 02 00 00 bb 03 00 00 00 48 83 05 d0 4e b4 02 01 4d 89 e5 48 8b 7d b8 e8 74 5d 87 00 48 83 05 ec 4e b4 02 01 c7 45 c0 00 00 00 00 <49> 8b 14 24 48 89 d0 48 c1 ea 35 48 8b 14 d5 00 00 a2 b0 48 c1
> [ 28.553655] RIP [<ffffffffadeed16b>] khugepaged_scan_pmd+0x33b/0xaa0
> [ 28.553659] RSP <ffff880035b7bd20>
> [ 28.553660] CR2: 0000000000000000
> [ 28.553666] ---[ end trace bb022a4fda57d1e5 ]---
> [ 28.553669] Kernel panic - not syncing: Fatal exception
Seems it should be something like that (untested at moment, please
take a look).
---
From: Cyrill Gorcunov <gorcunov@openvz.org>
Subject: [PATCH] mm: Fix tracepoint for scanning pages
The page might be NULL so use page_to_pfn in conditional call.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
---
include/trace/events/huge_memory.h | 12 ++++++------
mm/huge_memory.c | 6 +++---
2 files changed, 9 insertions(+), 9 deletions(-)
Index: linux-ml.git/include/trace/events/huge_memory.h
===================================================================
--- linux-ml.git.orig/include/trace/events/huge_memory.h
+++ linux-ml.git/include/trace/events/huge_memory.h
@@ -46,10 +46,10 @@ SCAN_STATUS
TRACE_EVENT(mm_khugepaged_scan_pmd,
- TP_PROTO(struct mm_struct *mm, unsigned long pfn, bool writable,
+ TP_PROTO(struct mm_struct *mm, struct page *page, bool writable,
bool referenced, int none_or_zero, int status),
- TP_ARGS(mm, pfn, writable, referenced, none_or_zero, status),
+ TP_ARGS(mm, page, writable, referenced, none_or_zero, status),
TP_STRUCT__entry(
__field(struct mm_struct *, mm)
@@ -62,7 +62,7 @@ TRACE_EVENT(mm_khugepaged_scan_pmd,
TP_fast_assign(
__entry->mm = mm;
- __entry->pfn = pfn;
+ __entry->pfn = page ? page_to_pfn(page) : -1UL;
__entry->writable = writable;
__entry->referenced = referenced;
__entry->none_or_zero = none_or_zero;
@@ -104,10 +104,10 @@ TRACE_EVENT(mm_collapse_huge_page,
TRACE_EVENT(mm_collapse_huge_page_isolate,
- TP_PROTO(unsigned long pfn, int none_or_zero,
+ TP_PROTO(struct page *page, int none_or_zero,
bool referenced, bool writable, int status),
- TP_ARGS(pfn, none_or_zero, referenced, writable, status),
+ TP_ARGS(page, none_or_zero, referenced, writable, status),
TP_STRUCT__entry(
__field(unsigned long, pfn)
@@ -118,7 +118,7 @@ TRACE_EVENT(mm_collapse_huge_page_isolat
),
TP_fast_assign(
- __entry->pfn = pfn;
+ __entry->pfn = page ? page_to_pfn(page) : -1UL;
__entry->none_or_zero = none_or_zero;
__entry->referenced = referenced;
__entry->writable = writable;
Index: linux-ml.git/mm/huge_memory.c
===================================================================
--- linux-ml.git.orig/mm/huge_memory.c
+++ linux-ml.git/mm/huge_memory.c
@@ -2068,7 +2068,7 @@ static int __collapse_huge_page_isolate(
if (likely(writable)) {
if (likely(referenced)) {
result = SCAN_SUCCEED;
- trace_mm_collapse_huge_page_isolate(page_to_pfn(page), none_or_zero,
+ trace_mm_collapse_huge_page_isolate(page, none_or_zero,
referenced, writable, result);
return 1;
}
@@ -2078,7 +2078,7 @@ static int __collapse_huge_page_isolate(
out:
release_pte_pages(pte, _pte);
- trace_mm_collapse_huge_page_isolate(page_to_pfn(page), none_or_zero,
+ trace_mm_collapse_huge_page_isolate(page, none_or_zero,
referenced, writable, result);
return 0;
}
@@ -2576,7 +2576,7 @@ out_unmap:
collapse_huge_page(mm, address, hpage, vma, node);
}
out:
- trace_mm_khugepaged_scan_pmd(mm, page_to_pfn(page), writable, referenced,
+ trace_mm_khugepaged_scan_pmd(mm, page, writable, referenced,
none_or_zero, result);
return ret;
}
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [lkp] [mm] 7d2eba0557: BUG: unable to handle kernel NULL pointer dereference at (null)
2016-01-19 8:02 ` [lkp] [mm] 7d2eba0557: BUG: unable to handle kernel NULL pointer dereference at (null) Cyrill Gorcunov
@ 2016-01-19 22:30 ` David Rientjes
2016-01-19 22:45 ` Cyrill Gorcunov
0 siblings, 1 reply; 5+ messages in thread
From: David Rientjes @ 2016-01-19 22:30 UTC (permalink / raw)
To: Cyrill Gorcunov
Cc: kernel test robot, Ebru Akagunduz, lkp, LKML, Andrew Morton,
Dan Carpenter, Michal Hocko, Johannes Weiner, Hugh Dickins,
Aneesh Kumar K.V, Vlastimil Babka, Mel Gorman, Xie XiuQi,
Joonsoo Kim, Andrea Arcangeli, Naoya Horiguchi, Rik van Riel,
Kirill A. Shutemov, Linus Torvalds
On Tue, 19 Jan 2016, Cyrill Gorcunov wrote:
> Seems it should be something like that (untested at moment, please
> take a look).
> ---
> From: Cyrill Gorcunov <gorcunov@openvz.org>
> Subject: [PATCH] mm: Fix tracepoint for scanning pages
>
> The page might be NULL so use page_to_pfn in conditional call.
>
> Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
I think this is obviously correct, so
Acked-by: David Rientjes <rientjes@google.com>
But we'd benefit from a Tested-by from the robot even more.
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [lkp] [mm] 7d2eba0557: BUG: unable to handle kernel NULL pointer dereference at (null)
2016-01-19 22:30 ` David Rientjes
@ 2016-01-19 22:45 ` Cyrill Gorcunov
2016-01-19 23:07 ` Vlastimil Babka
0 siblings, 1 reply; 5+ messages in thread
From: Cyrill Gorcunov @ 2016-01-19 22:45 UTC (permalink / raw)
To: David Rientjes
Cc: kernel test robot, Ebru Akagunduz, lkp, LKML, Andrew Morton,
Dan Carpenter, Michal Hocko, Johannes Weiner, Hugh Dickins,
Aneesh Kumar K.V, Vlastimil Babka, Mel Gorman, Xie XiuQi,
Joonsoo Kim, Andrea Arcangeli, Naoya Horiguchi, Rik van Riel,
Kirill A. Shutemov, Linus Torvalds
On Tue, Jan 19, 2016 at 02:30:45PM -0800, David Rientjes wrote:
> On Tue, 19 Jan 2016, Cyrill Gorcunov wrote:
>
> > Seems it should be something like that (untested at moment, please
> > take a look).
> > ---
> > From: Cyrill Gorcunov <gorcunov@openvz.org>
> > Subject: [PATCH] mm: Fix tracepoint for scanning pages
> >
> > The page might be NULL so use page_to_pfn in conditional call.
> >
> > Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
>
> I think this is obviously correct, so
>
> Acked-by: David Rientjes <rientjes@google.com>
Thank you. I've compile tested it, but didn't run.
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [lkp] [mm] 7d2eba0557: BUG: unable to handle kernel NULL pointer dereference at (null)
2016-01-19 22:45 ` Cyrill Gorcunov
@ 2016-01-19 23:07 ` Vlastimil Babka
2016-01-20 7:26 ` Cyrill Gorcunov
0 siblings, 1 reply; 5+ messages in thread
From: Vlastimil Babka @ 2016-01-19 23:07 UTC (permalink / raw)
To: Cyrill Gorcunov, David Rientjes, Andrew Morton
Cc: kernel test robot, Ebru Akagunduz, lkp, LKML, Dan Carpenter,
Michal Hocko, Johannes Weiner, Hugh Dickins, Aneesh Kumar K.V,
Mel Gorman, Xie XiuQi, Joonsoo Kim, Andrea Arcangeli,
Naoya Horiguchi, Rik van Riel, Kirill A. Shutemov, Linus Torvalds
On 01/19/2016 11:45 PM, Cyrill Gorcunov wrote:
> On Tue, Jan 19, 2016 at 02:30:45PM -0800, David Rientjes wrote:
>> On Tue, 19 Jan 2016, Cyrill Gorcunov wrote:
>>
>> > Seems it should be something like that (untested at moment, please
>> > take a look).
>> > ---
>> > From: Cyrill Gorcunov <gorcunov@openvz.org>
>> > Subject: [PATCH] mm: Fix tracepoint for scanning pages
>> >
>> > The page might be NULL so use page_to_pfn in conditional call.
>> >
>> > Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
>>
>> I think this is obviously correct, so
>>
>> Acked-by: David Rientjes <rientjes@google.com>
>
> Thank you. I've compile tested it, but didn't run.
This has been in mmotm since november:
http://ozlabs.org/~akpm/mmotm/broken-out/mm-fix-kernel-crash-in-khugepaged-thread.patch
Although your patch does use -1UL instead of -1 (AFAICS the only difference, but
it's too late here), which Steven Rostedt suggested, so it's probably better.
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [lkp] [mm] 7d2eba0557: BUG: unable to handle kernel NULL pointer dereference at (null)
2016-01-19 23:07 ` Vlastimil Babka
@ 2016-01-20 7:26 ` Cyrill Gorcunov
0 siblings, 0 replies; 5+ messages in thread
From: Cyrill Gorcunov @ 2016-01-20 7:26 UTC (permalink / raw)
To: Vlastimil Babka
Cc: David Rientjes, Andrew Morton, kernel test robot, Ebru Akagunduz,
lkp, LKML, Dan Carpenter, Michal Hocko, Johannes Weiner,
Hugh Dickins, Aneesh Kumar K.V, Mel Gorman, Xie XiuQi,
Joonsoo Kim, Andrea Arcangeli, Naoya Horiguchi, Rik van Riel,
Kirill A. Shutemov, Linus Torvalds
On Wed, Jan 20, 2016 at 12:07:46AM +0100, Vlastimil Babka wrote:
>
> This has been in mmotm since november:
> http://ozlabs.org/~akpm/mmotm/broken-out/mm-fix-kernel-crash-in-khugepaged-thread.patch
>
> Although your patch does use -1UL instead of -1 (AFAICS the only difference, but
> it's too late here), which Steven Rostedt suggested, so it's probably better.
OK, great then, drop mine.
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2016-01-20 7:26 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
[not found] <8737tue3mn.fsf@yhuang-dev.intel.com>
2016-01-19 8:02 ` [lkp] [mm] 7d2eba0557: BUG: unable to handle kernel NULL pointer dereference at (null) Cyrill Gorcunov
2016-01-19 22:30 ` David Rientjes
2016-01-19 22:45 ` Cyrill Gorcunov
2016-01-19 23:07 ` Vlastimil Babka
2016-01-20 7:26 ` Cyrill Gorcunov
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).