2018-08-29 07:49:29

by Larry Chen

[permalink] [raw]
Subject: [PATCH] fix crash on ocfs2_duplicate_clusters_by_page

ocfs2_duplicate_clusters_by_page may crash if one of extent's pages is dirty.
When a page has not been written back, it is still in dirty state. If
ocfs2_duplicate_clusters_by_page is called against the
dirty page, the crash happens.

To fix this bug, we can just unlock the page and wait the page until
it's not dirty.

The following is the buck trace dump:

kernel BUG at /root/code/ocfs2/refcounttree.c:2961!
[exception RIP: ocfs2_duplicate_clusters_by_page+822]
__ocfs2_move_extent+0x80/0x450 [ocfs2]
? __ocfs2_claim_clusters+0x130/0x250 [ocfs2]
ocfs2_defrag_extent+0x5b8/0x5e0 [ocfs2]
__ocfs2_move_extents_range+0x2a4/0x470 [ocfs2]
ocfs2_move_extents+0x180/0x3b0 [ocfs2]
? ocfs2_wait_for_recovery+0x13/0x70 [ocfs2]
ocfs2_ioctl_move_extents+0x133/0x2d0 [ocfs2]
ocfs2_ioctl+0x253/0x640 [ocfs2]
do_vfs_ioctl+0x90/0x5f0
SyS_ioctl+0x74/0x80
do_syscall_64+0x74/0x140
entry_SYSCALL_64_after_hwframe+0x3d/0xa2

Change-log:
1. Once we founce the page is dirty, we do not wait until it's clean,
but rather we use write_one_page to write it back

2. write_one_page arguments list changed, adjust and retest this patch.
Signed-off-by: Larry Chen <[email protected]>
---
fs/ocfs2/refcounttree.c | 12 ++++++++++--
1 file changed, 10 insertions(+), 2 deletions(-)

diff --git a/fs/ocfs2/refcounttree.c b/fs/ocfs2/refcounttree.c
index 7869622af22a..d6022e04c35a 100644
--- a/fs/ocfs2/refcounttree.c
+++ b/fs/ocfs2/refcounttree.c
@@ -2946,6 +2946,7 @@ int ocfs2_duplicate_clusters_by_page(handle_t *handle,
if (map_end & (PAGE_SIZE - 1))
to = map_end & (PAGE_SIZE - 1);

+retry:
page = find_or_create_page(mapping, page_index, GFP_NOFS);
if (!page) {
ret = -ENOMEM;
@@ -2957,8 +2958,15 @@ int ocfs2_duplicate_clusters_by_page(handle_t *handle,
* In case PAGE_SIZE <= CLUSTER_SIZE, This page
* can't be dirtied before we CoW it out.
*/
- if (PAGE_SIZE <= OCFS2_SB(sb)->s_clustersize)
- BUG_ON(PageDirty(page));
+ if (PAGE_SIZE <= OCFS2_SB(sb)->s_clustersize) {
+ if (PageDirty(page)) {
+ /*
+ * write_on_page will unlock the page on return
+ */
+ ret = write_one_page(page);
+ goto retry;
+ }
+ }

if (!PageUptodate(page)) {
ret = block_read_full_page(page, ocfs2_get_block);
--
2.13.7



2018-08-29 22:27:23

by Andrew Morton

[permalink] [raw]
Subject: Re: [PATCH] fix crash on ocfs2_duplicate_clusters_by_page

On Wed, 29 Aug 2018 15:47:40 +0800 Larry Chen <[email protected]> wrote:

> ocfs2_duplicate_clusters_by_page may crash if one of extent's pages is dirty.
> When a page has not been written back, it is still in dirty state. If
> ocfs2_duplicate_clusters_by_page is called against the
> dirty page, the crash happens.
>
> To fix this bug, we can just unlock the page and wait the page until
> it's not dirty.
>
> The following is the buck trace dump:
>
> kernel BUG at /root/code/ocfs2/refcounttree.c:2961!
> [exception RIP: ocfs2_duplicate_clusters_by_page+822]
> __ocfs2_move_extent+0x80/0x450 [ocfs2]
> ? __ocfs2_claim_clusters+0x130/0x250 [ocfs2]
> ocfs2_defrag_extent+0x5b8/0x5e0 [ocfs2]
> __ocfs2_move_extents_range+0x2a4/0x470 [ocfs2]
> ocfs2_move_extents+0x180/0x3b0 [ocfs2]
> ? ocfs2_wait_for_recovery+0x13/0x70 [ocfs2]
> ocfs2_ioctl_move_extents+0x133/0x2d0 [ocfs2]
> ocfs2_ioctl+0x253/0x640 [ocfs2]
> do_vfs_ioctl+0x90/0x5f0
> SyS_ioctl+0x74/0x80
> do_syscall_64+0x74/0x140
> entry_SYSCALL_64_after_hwframe+0x3d/0xa2
>
> --- a/fs/ocfs2/refcounttree.c
> +++ b/fs/ocfs2/refcounttree.c
> @@ -2946,6 +2946,7 @@ int ocfs2_duplicate_clusters_by_page(handle_t *handle,
> if (map_end & (PAGE_SIZE - 1))
> to = map_end & (PAGE_SIZE - 1);
>
> +retry:
> page = find_or_create_page(mapping, page_index, GFP_NOFS);
> if (!page) {
> ret = -ENOMEM;
> @@ -2957,8 +2958,15 @@ int ocfs2_duplicate_clusters_by_page(handle_t *handle,
> * In case PAGE_SIZE <= CLUSTER_SIZE, This page
> * can't be dirtied before we CoW it out.
> */

Looks sane, but the below change shows that the above comment is
untrue. Can we please update the comment as well?


> - if (PAGE_SIZE <= OCFS2_SB(sb)->s_clustersize)
> - BUG_ON(PageDirty(page));
> + if (PAGE_SIZE <= OCFS2_SB(sb)->s_clustersize) {
> + if (PageDirty(page)) {
> + /*
> + * write_on_page will unlock the page on return
> + */
> + ret = write_one_page(page);
> + goto retry;
> + }
> + }
>
> if (!PageUptodate(page)) {
> ret = block_read_full_page(page, ocfs2_get_block);


2018-08-30 02:32:26

by Changwei Ge

[permalink] [raw]
Subject: Re: [Ocfs2-devel] [PATCH] fix crash on ocfs2_duplicate_clusters_by_page

Hi Larry,

Besides Andrew's comments, other parts of this patch look good to me.

Thanks,

Changwei


On 2018/8/30 6:25, Andrew Morton wrote:
> On Wed, 29 Aug 2018 15:47:40 +0800 Larry Chen <[email protected]> wrote:
>
>> ocfs2_duplicate_clusters_by_page may crash if one of extent's pages is dirty.
>> When a page has not been written back, it is still in dirty state. If
>> ocfs2_duplicate_clusters_by_page is called against the
>> dirty page, the crash happens.
>>
>> To fix this bug, we can just unlock the page and wait the page until
>> it's not dirty.
>>
>> The following is the buck trace dump:
>>
>> kernel BUG at /root/code/ocfs2/refcounttree.c:2961!
>> [exception RIP: ocfs2_duplicate_clusters_by_page+822]
>> __ocfs2_move_extent+0x80/0x450 [ocfs2]
>> ? __ocfs2_claim_clusters+0x130/0x250 [ocfs2]
>> ocfs2_defrag_extent+0x5b8/0x5e0 [ocfs2]
>> __ocfs2_move_extents_range+0x2a4/0x470 [ocfs2]
>> ocfs2_move_extents+0x180/0x3b0 [ocfs2]
>> ? ocfs2_wait_for_recovery+0x13/0x70 [ocfs2]
>> ocfs2_ioctl_move_extents+0x133/0x2d0 [ocfs2]
>> ocfs2_ioctl+0x253/0x640 [ocfs2]
>> do_vfs_ioctl+0x90/0x5f0
>> SyS_ioctl+0x74/0x80
>> do_syscall_64+0x74/0x140
>> entry_SYSCALL_64_after_hwframe+0x3d/0xa2
>>
>> --- a/fs/ocfs2/refcounttree.c
>> +++ b/fs/ocfs2/refcounttree.c
>> @@ -2946,6 +2946,7 @@ int ocfs2_duplicate_clusters_by_page(handle_t *handle,
>> if (map_end & (PAGE_SIZE - 1))
>> to = map_end & (PAGE_SIZE - 1);
>>
>> +retry:
>> page = find_or_create_page(mapping, page_index, GFP_NOFS);
>> if (!page) {
>> ret = -ENOMEM;
>> @@ -2957,8 +2958,15 @@ int ocfs2_duplicate_clusters_by_page(handle_t *handle,
>> * In case PAGE_SIZE <= CLUSTER_SIZE, This page
>> * can't be dirtied before we CoW it out.
>> */
> Looks sane, but the below change shows that the above comment is
> untrue. Can we please update the comment as well?
>
>
>> - if (PAGE_SIZE <= OCFS2_SB(sb)->s_clustersize)
>> - BUG_ON(PageDirty(page));
>> + if (PAGE_SIZE <= OCFS2_SB(sb)->s_clustersize) {
>> + if (PageDirty(page)) {
>> + /*
>> + * write_on_page will unlock the page on return
>> + */
>> + ret = write_one_page(page);
>> + goto retry;
>> + }
>> + }
>>
>> if (!PageUptodate(page)) {
>> ret = block_read_full_page(page, ocfs2_get_block);
>
> _______________________________________________
> Ocfs2-devel mailing list
> [email protected]
> https://oss.oracle.com/mailman/listinfo/ocfs2-devel

2018-08-30 02:43:02

by Larry Chen

[permalink] [raw]
Subject: Re: [Ocfs2-devel] [PATCH] fix crash on ocfs2_duplicate_clusters_by_page

Hi Andrew and Changwei,

Thanks for your review, I'll propose a new version with comments corrected.

Thanks,
Larry
On 08/30/2018 10:29 AM, Changwei Ge wrote:
> Hi Larry,
>
> Besides Andrew's comments, other parts of this patch look good to me.
>
> Thanks,
>
> Changwei
>
>
> On 2018/8/30 6:25, Andrew Morton wrote:
>> On Wed, 29 Aug 2018 15:47:40 +0800 Larry Chen <[email protected]> wrote:
>>
>>> ocfs2_duplicate_clusters_by_page may crash if one of extent's pages is dirty.
>>> When a page has not been written back, it is still in dirty state. If
>>> ocfs2_duplicate_clusters_by_page is called against the
>>> dirty page, the crash happens.
>>>
>>> To fix this bug, we can just unlock the page and wait the page until
>>> it's not dirty.
>>>
>>> The following is the buck trace dump:
>>>
>>> kernel BUG at /root/code/ocfs2/refcounttree.c:2961!
>>> [exception RIP: ocfs2_duplicate_clusters_by_page+822]
>>> __ocfs2_move_extent+0x80/0x450 [ocfs2]
>>> ? __ocfs2_claim_clusters+0x130/0x250 [ocfs2]
>>> ocfs2_defrag_extent+0x5b8/0x5e0 [ocfs2]
>>> __ocfs2_move_extents_range+0x2a4/0x470 [ocfs2]
>>> ocfs2_move_extents+0x180/0x3b0 [ocfs2]
>>> ? ocfs2_wait_for_recovery+0x13/0x70 [ocfs2]
>>> ocfs2_ioctl_move_extents+0x133/0x2d0 [ocfs2]
>>> ocfs2_ioctl+0x253/0x640 [ocfs2]
>>> do_vfs_ioctl+0x90/0x5f0
>>> SyS_ioctl+0x74/0x80
>>> do_syscall_64+0x74/0x140
>>> entry_SYSCALL_64_after_hwframe+0x3d/0xa2
>>>
>>> --- a/fs/ocfs2/refcounttree.c
>>> +++ b/fs/ocfs2/refcounttree.c
>>> @@ -2946,6 +2946,7 @@ int ocfs2_duplicate_clusters_by_page(handle_t *handle,
>>> if (map_end & (PAGE_SIZE - 1))
>>> to = map_end & (PAGE_SIZE - 1);
>>>
>>> +retry:
>>> page = find_or_create_page(mapping, page_index, GFP_NOFS);
>>> if (!page) {
>>> ret = -ENOMEM;
>>> @@ -2957,8 +2958,15 @@ int ocfs2_duplicate_clusters_by_page(handle_t *handle,
>>> * In case PAGE_SIZE <= CLUSTER_SIZE, This page
>>> * can't be dirtied before we CoW it out.
>>> */
>> Looks sane, but the below change shows that the above comment is
>> untrue. Can we please update the comment as well?
>>
>>
>>> - if (PAGE_SIZE <= OCFS2_SB(sb)->s_clustersize)
>>> - BUG_ON(PageDirty(page));
>>> + if (PAGE_SIZE <= OCFS2_SB(sb)->s_clustersize) {
>>> + if (PageDirty(page)) {
>>> + /*
>>> + * write_on_page will unlock the page on return
>>> + */
>>> + ret = write_one_page(page);
>>> + goto retry;
>>> + }
>>> + }
>>>
>>> if (!PageUptodate(page)) {
>>> ret = block_read_full_page(page, ocfs2_get_block);
>>
>> _______________________________________________
>> Ocfs2-devel mailing list
>> [email protected]
>> https://oss.oracle.com/mailman/listinfo/ocfs2-devel
>