* + mm-migrate-fix-ref-count-handling-when-hugepage_migration_supported-v2.patch added to -mm tree
@ 2017-05-25 22:46 akpm
2017-05-26 5:41 ` Naoya Horiguchi
` (2 more replies)
0 siblings, 3 replies; 6+ messages in thread
From: akpm @ 2017-05-25 22:46 UTC (permalink / raw)
To: punit.agrawal, cl, iamjoonsoo.kim, manoj.iyer, mgorman,
n-horiguchi, stable, wanpeng.li, mm-commits
The patch titled
Subject: mm/migrate: fix refcount handling when !hugepage_migration_supported()
has been added to the -mm tree. Its filename is
mm-migrate-fix-ref-count-handling-when-hugepage_migration_supported-v2.patch
This patch should soon appear at
http://ozlabs.org/~akpm/mmots/broken-out/mm-migrate-fix-ref-count-handling-when-hugepage_migration_supported-v2.patch
and later at
http://ozlabs.org/~akpm/mmotm/broken-out/mm-migrate-fix-ref-count-handling-when-hugepage_migration_supported-v2.patch
Before you just go and hit "reply", please:
a) Consider who else should be cc'ed
b) Prefer to cc a suitable mailing list as well
c) Ideally: find the original patch on the mailing list and do a
reply-to-all to that, adding suitable additional cc's
*** Remember to use Documentation/SubmitChecklist when testing your code ***
The -mm tree is included into linux-next and is updated
there every 3-4 working days
------------------------------------------------------
From: Punit Agrawal <punit.agrawal@arm.com>
Subject: mm/migrate: fix refcount handling when !hugepage_migration_supported()
On failing to migrate a page, soft_offline_huge_page() performs the
necessary update to the hugepage ref-count. When
!hugepage_migration_supported() , unmap_and_move_hugepage() also
decrements the page ref-count for the hugepage. The combined behaviour
leaves the ref-count in an inconsistent state.
This leads to soft lockups when running the overcommitted hugepage test
from mce-tests suite.
Soft offlining pfn 0x83ed600 at process virtual address 0x400000000000
soft offline: 0x83ed600: migration failed 1, type
1fffc00000008008 (uptodate|head)
INFO: rcu_preempt detected stalls on CPUs/tasks:
Tasks blocked on level-0 rcu_node (CPUs 0-7): P2715
(detected by 7, t=5254 jiffies, g=963, c=962, q=321)
thugetlb_overco R running task 0 2715 2685 0x00000008
Call trace:
[<ffff000008089f90>] dump_backtrace+0x0/0x268
[<ffff00000808a2d4>] show_stack+0x24/0x30
[<ffff000008100d34>] sched_show_task+0x134/0x180
[<ffff0000081c90fc>] rcu_print_detail_task_stall_rnp+0x54/0x7c
[<ffff00000813cfd4>] rcu_check_callbacks+0xa74/0xb08
[<ffff000008143a3c>] update_process_times+0x34/0x60
[<ffff0000081550e8>] tick_sched_handle.isra.7+0x38/0x70
[<ffff00000815516c>] tick_sched_timer+0x4c/0x98
[<ffff0000081442e0>] __hrtimer_run_queues+0xc0/0x300
[<ffff000008144fa4>] hrtimer_interrupt+0xac/0x228
[<ffff0000089a56d4>] arch_timer_handler_phys+0x3c/0x50
[<ffff00000812f1bc>] handle_percpu_devid_irq+0x8c/0x290
[<ffff0000081297fc>] generic_handle_irq+0x34/0x50
[<ffff000008129f00>] __handle_domain_irq+0x68/0xc0
[<ffff0000080816b4>] gic_handle_irq+0x5c/0xb0
Address this by changing the putback_active_hugepage() in
soft_offline_huge_page() to putback_movable_pages().
Link: http://lkml.kernel.org/r/20170525135146.32011-1-punit.agrawal@arm.com
Reported-by: Manoj Iyer <manoj.iyer@canonical.com>
Tested-by: Manoj Iyer <manoj.iyer@canonical.com>
Suggested-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Signed-off-by: Punit Agrawal <punit.agrawal@arm.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Wanpeng Li <wanpeng.li@hotmail.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: <stable@vger.kernel.org> [3.14+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---
mm/memory-failure.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff -puN mm/memory-failure.c~mm-migrate-fix-ref-count-handling-when-hugepage_migration_supported-v2 mm/memory-failure.c
--- a/mm/memory-failure.c~mm-migrate-fix-ref-count-handling-when-hugepage_migration_supported-v2
+++ a/mm/memory-failure.c
@@ -1600,7 +1600,8 @@ static int soft_offline_huge_page(struct
* only one hugepage pointed to by hpage, so we need not
* run through the pagelist here.
*/
- putback_active_hugepage(hpage);
+ if (!list_empty(&pagelist))
+ putback_movable_pages(&pagelist);
if (ret > 0)
ret = -EIO;
} else {
_
Patches currently in -mm which might be from punit.agrawal@arm.com are
mm-migrate-fix-ref-count-handling-when-hugepage_migration_supported-v2.patch
mm-gup-ensure-real-head-page-is-ref-counted-when-using-hugepages.patch
mm-hugetlb-add-size-parameter-to-huge_pte_offset.patch
mm-hugetlb-allow-architectures-to-override-huge_pte_clear.patch
mm-hugetlb-introduce-set_huge_swap_pte_at-helper.patch
mm-hugetlb-introduce-set_huge_swap_pte_at-helper-v4.patch
mm-hugetlb-introduce-set_huge_swap_pte_at-helper-v41.patch
mm-rmap-use-correct-helper-when-poisoning-hugepages.patch
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: + mm-migrate-fix-ref-count-handling-when-hugepage_migration_supported-v2.patch added to -mm tree
2017-05-25 22:46 + mm-migrate-fix-ref-count-handling-when-hugepage_migration_supported-v2.patch added to -mm tree akpm
@ 2017-05-26 5:41 ` Naoya Horiguchi
2017-05-26 9:22 ` Punit Agrawal
2017-05-26 10:42 ` Mel Gorman
2 siblings, 0 replies; 6+ messages in thread
From: Naoya Horiguchi @ 2017-05-26 5:41 UTC (permalink / raw)
To: akpm
Cc: punit.agrawal, cl, iamjoonsoo.kim, manoj.iyer, mgorman, stable,
wanpeng.li, mm-commits
On Thu, May 25, 2017 at 03:46:23PM -0700, akpm@linux-foundation.org wrote:
>
> The patch titled
> Subject: mm/migrate: fix refcount handling when !hugepage_migration_supported()
> has been added to the -mm tree. Its filename is
> mm-migrate-fix-ref-count-handling-when-hugepage_migration_supported-v2.patch
>
> This patch should soon appear at
> http://ozlabs.org/~akpm/mmots/broken-out/mm-migrate-fix-ref-count-handling-when-hugepage_migration_supported-v2.patch
> and later at
> http://ozlabs.org/~akpm/mmotm/broken-out/mm-migrate-fix-ref-count-handling-when-hugepage_migration_supported-v2.patch
>
> Before you just go and hit "reply", please:
> a) Consider who else should be cc'ed
> b) Prefer to cc a suitable mailing list as well
> c) Ideally: find the original patch on the mailing list and do a
> reply-to-all to that, adding suitable additional cc's
>
> *** Remember to use Documentation/SubmitChecklist when testing your code ***
>
> The -mm tree is included into linux-next and is updated
> there every 3-4 working days
>
> ------------------------------------------------------
> From: Punit Agrawal <punit.agrawal@arm.com>
> Subject: mm/migrate: fix refcount handling when !hugepage_migration_supported()
>
> On failing to migrate a page, soft_offline_huge_page() performs the
> necessary update to the hugepage ref-count. When
> !hugepage_migration_supported() , unmap_and_move_hugepage() also
> decrements the page ref-count for the hugepage. The combined behaviour
> leaves the ref-count in an inconsistent state.
>
> This leads to soft lockups when running the overcommitted hugepage test
> from mce-tests suite.
>
> Soft offlining pfn 0x83ed600 at process virtual address 0x400000000000
> soft offline: 0x83ed600: migration failed 1, type
> 1fffc00000008008 (uptodate|head)
> INFO: rcu_preempt detected stalls on CPUs/tasks:
> Tasks blocked on level-0 rcu_node (CPUs 0-7): P2715
> (detected by 7, t=5254 jiffies, g=963, c=962, q=321)
> thugetlb_overco R running task 0 2715 2685 0x00000008
> Call trace:
> [<ffff000008089f90>] dump_backtrace+0x0/0x268
> [<ffff00000808a2d4>] show_stack+0x24/0x30
> [<ffff000008100d34>] sched_show_task+0x134/0x180
> [<ffff0000081c90fc>] rcu_print_detail_task_stall_rnp+0x54/0x7c
> [<ffff00000813cfd4>] rcu_check_callbacks+0xa74/0xb08
> [<ffff000008143a3c>] update_process_times+0x34/0x60
> [<ffff0000081550e8>] tick_sched_handle.isra.7+0x38/0x70
> [<ffff00000815516c>] tick_sched_timer+0x4c/0x98
> [<ffff0000081442e0>] __hrtimer_run_queues+0xc0/0x300
> [<ffff000008144fa4>] hrtimer_interrupt+0xac/0x228
> [<ffff0000089a56d4>] arch_timer_handler_phys+0x3c/0x50
> [<ffff00000812f1bc>] handle_percpu_devid_irq+0x8c/0x290
> [<ffff0000081297fc>] generic_handle_irq+0x34/0x50
> [<ffff000008129f00>] __handle_domain_irq+0x68/0xc0
> [<ffff0000080816b4>] gic_handle_irq+0x5c/0xb0
>
> Address this by changing the putback_active_hugepage() in
> soft_offline_huge_page() to putback_movable_pages().
>
> Link: http://lkml.kernel.org/r/20170525135146.32011-1-punit.agrawal@arm.com
> Reported-by: Manoj Iyer <manoj.iyer@canonical.com>
> Tested-by: Manoj Iyer <manoj.iyer@canonical.com>
> Suggested-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
> Signed-off-by: Punit Agrawal <punit.agrawal@arm.com>
> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
> Cc: Wanpeng Li <wanpeng.li@hotmail.com>
> Cc: Christoph Lameter <cl@linux.com>
> Cc: Mel Gorman <mgorman@techsingularity.net>
> Cc: <stable@vger.kernel.org> [3.14+]
> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
> ---
>
> mm/memory-failure.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff -puN mm/memory-failure.c~mm-migrate-fix-ref-count-handling-when-hugepage_migration_supported-v2 mm/memory-failure.c
> --- a/mm/memory-failure.c~mm-migrate-fix-ref-count-handling-when-hugepage_migration_supported-v2
> +++ a/mm/memory-failure.c
> @@ -1600,7 +1600,8 @@ static int soft_offline_huge_page(struct
> * only one hugepage pointed to by hpage, so we need not
> * run through the pagelist here.
> */
Hi Andrew,
I just noticed that the comment above this becomes obsolete with this change,
so could you remove it? (please fold it to this patch.)
@@ -1649,11 +1649,6 @@ static int soft_offline_huge_page(struct page *page, int flags)
if (ret) {
pr_info("soft offline: %#lx: migration failed %d, type %lx (%pGp)\n",
pfn, ret, page->flags, &page->flags);
- /*
- * We know that soft_offline_huge_page() tries to migrate
- * only one hugepage pointed to by hpage, so we need not
- * run through the pagelist here.
- */
if (!list_empty(&pagelist))
putback_movable_pages(&pagelist);
if (ret > 0)
Thanks,
Naoya Horiguchi
> - putback_active_hugepage(hpage);
> + if (!list_empty(&pagelist))
> + putback_movable_pages(&pagelist);
> if (ret > 0)
> ret = -EIO;
> } else {
> _
>
> Patches currently in -mm which might be from punit.agrawal@arm.com are
>
> mm-migrate-fix-ref-count-handling-when-hugepage_migration_supported-v2.patch
> mm-gup-ensure-real-head-page-is-ref-counted-when-using-hugepages.patch
> mm-hugetlb-add-size-parameter-to-huge_pte_offset.patch
> mm-hugetlb-allow-architectures-to-override-huge_pte_clear.patch
> mm-hugetlb-introduce-set_huge_swap_pte_at-helper.patch
> mm-hugetlb-introduce-set_huge_swap_pte_at-helper-v4.patch
> mm-hugetlb-introduce-set_huge_swap_pte_at-helper-v41.patch
> mm-rmap-use-correct-helper-when-poisoning-hugepages.patch
>
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: + mm-migrate-fix-ref-count-handling-when-hugepage_migration_supported-v2.patch added to -mm tree
2017-05-25 22:46 + mm-migrate-fix-ref-count-handling-when-hugepage_migration_supported-v2.patch added to -mm tree akpm
2017-05-26 5:41 ` Naoya Horiguchi
@ 2017-05-26 9:22 ` Punit Agrawal
2017-05-26 10:42 ` Mel Gorman
2 siblings, 0 replies; 6+ messages in thread
From: Punit Agrawal @ 2017-05-26 9:22 UTC (permalink / raw)
To: akpm
Cc: cl, iamjoonsoo.kim, manoj.iyer, mgorman, n-horiguchi, stable,
wanpeng.li, mm-commits
<akpm@linux-foundation.org> writes:
> The patch titled
> Subject: mm/migrate: fix refcount handling when !hugepage_migration_supported()
> has been added to the -mm tree. Its filename is
> mm-migrate-fix-ref-count-handling-when-hugepage_migration_supported-v2.patch
>
> This patch should soon appear at
> http://ozlabs.org/~akpm/mmots/broken-out/mm-migrate-fix-ref-count-handling-when-hugepage_migration_supported-v2.patch
> and later at
> http://ozlabs.org/~akpm/mmotm/broken-out/mm-migrate-fix-ref-count-handling-when-hugepage_migration_supported-v2.patch
>
> Before you just go and hit "reply", please:
> a) Consider who else should be cc'ed
> b) Prefer to cc a suitable mailing list as well
> c) Ideally: find the original patch on the mailing list and do a
> reply-to-all to that, adding suitable additional cc's
>
> *** Remember to use Documentation/SubmitChecklist when testing your code ***
>
> The -mm tree is included into linux-next and is updated
> there every 3-4 working days
>
> ------------------------------------------------------
> From: Punit Agrawal <punit.agrawal@arm.com>
> Subject: mm/migrate: fix refcount handling when !hugepage_migration_supported()
>
> On failing to migrate a page, soft_offline_huge_page() performs the
> necessary update to the hugepage ref-count. When
> !hugepage_migration_supported() , unmap_and_move_hugepage() also
> decrements the page ref-count for the hugepage. The combined behaviour
> leaves the ref-count in an inconsistent state.
>
> This leads to soft lockups when running the overcommitted hugepage test
> from mce-tests suite.
>
> Soft offlining pfn 0x83ed600 at process virtual address 0x400000000000
> soft offline: 0x83ed600: migration failed 1, type
> 1fffc00000008008 (uptodate|head)
> INFO: rcu_preempt detected stalls on CPUs/tasks:
> Tasks blocked on level-0 rcu_node (CPUs 0-7): P2715
> (detected by 7, t=5254 jiffies, g=963, c=962, q=321)
> thugetlb_overco R running task 0 2715 2685 0x00000008
> Call trace:
> [<ffff000008089f90>] dump_backtrace+0x0/0x268
> [<ffff00000808a2d4>] show_stack+0x24/0x30
> [<ffff000008100d34>] sched_show_task+0x134/0x180
> [<ffff0000081c90fc>] rcu_print_detail_task_stall_rnp+0x54/0x7c
> [<ffff00000813cfd4>] rcu_check_callbacks+0xa74/0xb08
> [<ffff000008143a3c>] update_process_times+0x34/0x60
> [<ffff0000081550e8>] tick_sched_handle.isra.7+0x38/0x70
> [<ffff00000815516c>] tick_sched_timer+0x4c/0x98
> [<ffff0000081442e0>] __hrtimer_run_queues+0xc0/0x300
> [<ffff000008144fa4>] hrtimer_interrupt+0xac/0x228
> [<ffff0000089a56d4>] arch_timer_handler_phys+0x3c/0x50
> [<ffff00000812f1bc>] handle_percpu_devid_irq+0x8c/0x290
> [<ffff0000081297fc>] generic_handle_irq+0x34/0x50
> [<ffff000008129f00>] __handle_domain_irq+0x68/0xc0
> [<ffff0000080816b4>] gic_handle_irq+0x5c/0xb0
>
> Address this by changing the putback_active_hugepage() in
> soft_offline_huge_page() to putback_movable_pages().
>
> Link: http://lkml.kernel.org/r/20170525135146.32011-1-punit.agrawal@arm.com
> Reported-by: Manoj Iyer <manoj.iyer@canonical.com>
> Tested-by: Manoj Iyer <manoj.iyer@canonical.com>
> Suggested-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
> Signed-off-by: Punit Agrawal <punit.agrawal@arm.com>
> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
> Cc: Wanpeng Li <wanpeng.li@hotmail.com>
> Cc: Christoph Lameter <cl@linux.com>
> Cc: Mel Gorman <mgorman@techsingularity.net>
> Cc: <stable@vger.kernel.org> [3.14+]
> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Hi Andrew,
Thanks a lot for fixing up the stable email address. I need to introduce
some form of sanity checking of emails in my flow considering the number
of bounces I am generating with the current set of patches. Manual
inspection just isn't cutting it.
Appreciate you picking up all the patches.
Thanks,
Punit
> ---
>
> mm/memory-failure.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff -puN mm/memory-failure.c~mm-migrate-fix-ref-count-handling-when-hugepage_migration_supported-v2 mm/memory-failure.c
> --- a/mm/memory-failure.c~mm-migrate-fix-ref-count-handling-when-hugepage_migration_supported-v2
> +++ a/mm/memory-failure.c
> @@ -1600,7 +1600,8 @@ static int soft_offline_huge_page(struct
> * only one hugepage pointed to by hpage, so we need not
> * run through the pagelist here.
> */
> - putback_active_hugepage(hpage);
> + if (!list_empty(&pagelist))
> + putback_movable_pages(&pagelist);
> if (ret > 0)
> ret = -EIO;
> } else {
> _
>
> Patches currently in -mm which might be from punit.agrawal@arm.com are
>
> mm-migrate-fix-ref-count-handling-when-hugepage_migration_supported-v2.patch
> mm-gup-ensure-real-head-page-is-ref-counted-when-using-hugepages.patch
> mm-hugetlb-add-size-parameter-to-huge_pte_offset.patch
> mm-hugetlb-allow-architectures-to-override-huge_pte_clear.patch
> mm-hugetlb-introduce-set_huge_swap_pte_at-helper.patch
> mm-hugetlb-introduce-set_huge_swap_pte_at-helper-v4.patch
> mm-hugetlb-introduce-set_huge_swap_pte_at-helper-v41.patch
> mm-rmap-use-correct-helper-when-poisoning-hugepages.patch
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: + mm-migrate-fix-ref-count-handling-when-hugepage_migration_supported-v2.patch added to -mm tree
2017-05-25 22:46 + mm-migrate-fix-ref-count-handling-when-hugepage_migration_supported-v2.patch added to -mm tree akpm
2017-05-26 5:41 ` Naoya Horiguchi
2017-05-26 9:22 ` Punit Agrawal
@ 2017-05-26 10:42 ` Mel Gorman
2017-05-26 14:52 ` Punit Agrawal
2 siblings, 1 reply; 6+ messages in thread
From: Mel Gorman @ 2017-05-26 10:42 UTC (permalink / raw)
To: akpm
Cc: punit.agrawal, cl, iamjoonsoo.kim, manoj.iyer, n-horiguchi,
stable, wanpeng.li, mm-commits
On Thu, May 25, 2017 at 03:46:23PM -0700, akpm@linux-foundation.org wrote:
> ------------------------------------------------------
> From: Punit Agrawal <punit.agrawal@arm.com>
> Subject: mm/migrate: fix refcount handling when !hugepage_migration_supported()
>
> On failing to migrate a page, soft_offline_huge_page() performs the
> necessary update to the hugepage ref-count. When
> !hugepage_migration_supported() , unmap_and_move_hugepage() also
> decrements the page ref-count for the hugepage. The combined behaviour
> leaves the ref-count in an inconsistent state.
>
> This leads to soft lockups when running the overcommitted hugepage test
> from mce-tests suite.
>
> Soft offlining pfn 0x83ed600 at process virtual address 0x400000000000
> soft offline: 0x83ed600: migration failed 1, type
> 1fffc00000008008 (uptodate|head)
> INFO: rcu_preempt detected stalls on CPUs/tasks:
> Tasks blocked on level-0 rcu_node (CPUs 0-7): P2715
> (detected by 7, t=5254 jiffies, g=963, c=962, q=321)
> thugetlb_overco R running task 0 2715 2685 0x00000008
> Call trace:
> [<ffff000008089f90>] dump_backtrace+0x0/0x268
> [<ffff00000808a2d4>] show_stack+0x24/0x30
> [<ffff000008100d34>] sched_show_task+0x134/0x180
> [<ffff0000081c90fc>] rcu_print_detail_task_stall_rnp+0x54/0x7c
> [<ffff00000813cfd4>] rcu_check_callbacks+0xa74/0xb08
> [<ffff000008143a3c>] update_process_times+0x34/0x60
> [<ffff0000081550e8>] tick_sched_handle.isra.7+0x38/0x70
> [<ffff00000815516c>] tick_sched_timer+0x4c/0x98
> [<ffff0000081442e0>] __hrtimer_run_queues+0xc0/0x300
> [<ffff000008144fa4>] hrtimer_interrupt+0xac/0x228
> [<ffff0000089a56d4>] arch_timer_handler_phys+0x3c/0x50
> [<ffff00000812f1bc>] handle_percpu_devid_irq+0x8c/0x290
> [<ffff0000081297fc>] generic_handle_irq+0x34/0x50
> [<ffff000008129f00>] __handle_domain_irq+0x68/0xc0
> [<ffff0000080816b4>] gic_handle_irq+0x5c/0xb0
>
> Address this by changing the putback_active_hugepage() in
> soft_offline_huge_page() to putback_movable_pages().
>
I've never looked too closely at how hardware poisoning and hugetlb pages
migration is handled so I could easily have missed something but this
changelog and patch confuses me.
Surely if the inconsistency is between hugepage_migration_supported and
!hugepage_migration_supported then the check in soft_offline_huge_page()
should also be related to hugepage_migration_supported either in
soft_offline_huge_page() or in putback_movable_pages()?
--
Mel Gorman
SUSE Labs
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: + mm-migrate-fix-ref-count-handling-when-hugepage_migration_supported-v2.patch added to -mm tree
2017-05-26 10:42 ` Mel Gorman
@ 2017-05-26 14:52 ` Punit Agrawal
2017-05-26 15:33 ` Mel Gorman
0 siblings, 1 reply; 6+ messages in thread
From: Punit Agrawal @ 2017-05-26 14:52 UTC (permalink / raw)
To: Mel Gorman
Cc: akpm, cl, iamjoonsoo.kim, manoj.iyer, n-horiguchi, stable,
wanpeng.li, mm-commits
Mel Gorman <mgorman@techsingularity.net> writes:
> On Thu, May 25, 2017 at 03:46:23PM -0700, akpm@linux-foundation.org wrote:
>> ------------------------------------------------------
>> From: Punit Agrawal <punit.agrawal@arm.com>
>> Subject: mm/migrate: fix refcount handling when !hugepage_migration_supported()
>>
>> On failing to migrate a page, soft_offline_huge_page() performs the
>> necessary update to the hugepage ref-count. When
>> !hugepage_migration_supported() , unmap_and_move_hugepage() also
>> decrements the page ref-count for the hugepage. The combined behaviour
>> leaves the ref-count in an inconsistent state.
>>
>> This leads to soft lockups when running the overcommitted hugepage test
>> from mce-tests suite.
>>
>> Soft offlining pfn 0x83ed600 at process virtual address 0x400000000000
>> soft offline: 0x83ed600: migration failed 1, type
>> 1fffc00000008008 (uptodate|head)
>> INFO: rcu_preempt detected stalls on CPUs/tasks:
>> Tasks blocked on level-0 rcu_node (CPUs 0-7): P2715
>> (detected by 7, t=5254 jiffies, g=963, c=962, q=321)
>> thugetlb_overco R running task 0 2715 2685 0x00000008
>> Call trace:
>> [<ffff000008089f90>] dump_backtrace+0x0/0x268
>> [<ffff00000808a2d4>] show_stack+0x24/0x30
>> [<ffff000008100d34>] sched_show_task+0x134/0x180
>> [<ffff0000081c90fc>] rcu_print_detail_task_stall_rnp+0x54/0x7c
>> [<ffff00000813cfd4>] rcu_check_callbacks+0xa74/0xb08
>> [<ffff000008143a3c>] update_process_times+0x34/0x60
>> [<ffff0000081550e8>] tick_sched_handle.isra.7+0x38/0x70
>> [<ffff00000815516c>] tick_sched_timer+0x4c/0x98
>> [<ffff0000081442e0>] __hrtimer_run_queues+0xc0/0x300
>> [<ffff000008144fa4>] hrtimer_interrupt+0xac/0x228
>> [<ffff0000089a56d4>] arch_timer_handler_phys+0x3c/0x50
>> [<ffff00000812f1bc>] handle_percpu_devid_irq+0x8c/0x290
>> [<ffff0000081297fc>] generic_handle_irq+0x34/0x50
>> [<ffff000008129f00>] __handle_domain_irq+0x68/0xc0
>> [<ffff0000080816b4>] gic_handle_irq+0x5c/0xb0
>>
>> Address this by changing the putback_active_hugepage() in
>> soft_offline_huge_page() to putback_movable_pages().
>>
>
> I've never looked too closely at how hardware poisoning and hugetlb pages
> migration is handled so I could easily have missed something but this
> changelog and patch confuses me.
>
> Surely if the inconsistency is between hugepage_migration_supported and
> !hugepage_migration_supported then the check in soft_offline_huge_page()
> should also be related to hugepage_migration_supported either in
> soft_offline_huge_page() or in putback_movable_pages()?
The first version of the patch did indeed make a change that was around
!hugepage_migration_supported() [0] which was effectively a revert of
32665f2bbfed ("mm/migrate: correct failure handling if
!hugepage_migration_support()").
But Horiguchi-san suggested that dropping the putback_active_hugepage()
from unmap_and_move_hugepage() will bring back the issue that
32665f2bbfed addressed it was safer to take the current approach. It
also matches the pattern followed for !hugepage.
I did update the changelog but perhaps not enough - would updating the
changelog to reflect this help make it clearer?
Thanks,
Punit
[0] https://www.spinics.net/lists/kernel/msg2516334.html
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: + mm-migrate-fix-ref-count-handling-when-hugepage_migration_supported-v2.patch added to -mm tree
2017-05-26 14:52 ` Punit Agrawal
@ 2017-05-26 15:33 ` Mel Gorman
0 siblings, 0 replies; 6+ messages in thread
From: Mel Gorman @ 2017-05-26 15:33 UTC (permalink / raw)
To: Punit Agrawal
Cc: akpm, cl, iamjoonsoo.kim, manoj.iyer, n-horiguchi, stable,
wanpeng.li, mm-commits
On Fri, May 26, 2017 at 03:52:21PM +0100, Punit Agrawal wrote:
> > I've never looked too closely at how hardware poisoning and hugetlb pages
> > migration is handled so I could easily have missed something but this
> > changelog and patch confuses me.
> >
> > Surely if the inconsistency is between hugepage_migration_supported and
> > !hugepage_migration_supported then the check in soft_offline_huge_page()
> > should also be related to hugepage_migration_supported either in
> > soft_offline_huge_page() or in putback_movable_pages()?
>
> The first version of the patch did indeed make a change that was around
> !hugepage_migration_supported() [0] which was effectively a revert of
> 32665f2bbfed ("mm/migrate: correct failure handling if
> !hugepage_migration_support()").
>
> But Horiguchi-san suggested that dropping the putback_active_hugepage()
> from unmap_and_move_hugepage() will bring back the issue that
> 32665f2bbfed addressed it was safer to take the current approach. It
> also matches the pattern followed for !hugepage.
>
> I did update the changelog but perhaps not enough - would updating the
> changelog to reflect this help make it clearer?
>
Ok, I see what the patch is doing now. I don't think you need to alter
the changelog as I think the patch is ok.
Thanks for clarifying.
--
Mel Gorman
SUSE Labs
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2017-05-26 15:41 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-05-25 22:46 + mm-migrate-fix-ref-count-handling-when-hugepage_migration_supported-v2.patch added to -mm tree akpm
2017-05-26 5:41 ` Naoya Horiguchi
2017-05-26 9:22 ` Punit Agrawal
2017-05-26 10:42 ` Mel Gorman
2017-05-26 14:52 ` Punit Agrawal
2017-05-26 15:33 ` Mel Gorman
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.