Chao has pointed out that, maybe writebacking dentry page can redirty
dnode page
which may be clean previously, so it is more safe to call
f2fs_balance_fs to make sure
dirty metadatas of filesystem will not exceed the threshold.
On 2017/8/4 11:34, Jaegeuk Kim wrote:
On 08/04, Yunlong Song
Chao has pointed out that, maybe writebacking dentry page can redirty
dnode page
which may be clean previously, so it is more safe to call
f2fs_balance_fs to make sure
dirty metadatas of filesystem will not exceed the threshold.
On 2017/8/4 11:34, Jaegeuk Kim wrote:
On 08/04, Yunlong Song
On 08/04, Yunlong Song wrote:
> Since __write_data_page will not do f2fs_balance_fs for dir inode, so there
> is no lock.
> - f2fs_balance_fs
> - __write_data_page (dir inode)
> - f2fs_balance_fs again?<- Can not happen!
> And if let sync_dirty_inodes flush dentry page of inodeB,
On 08/04, Yunlong Song wrote:
> Since __write_data_page will not do f2fs_balance_fs for dir inode, so there
> is no lock.
> - f2fs_balance_fs
> - __write_data_page (dir inode)
> - f2fs_balance_fs again?<- Can not happen!
> And if let sync_dirty_inodes flush dentry page of inodeB,
Since __write_data_page will not do f2fs_balance_fs for dir inode, so
there is no lock.
- f2fs_balance_fs
- __write_data_page (dir inode)
- f2fs_balance_fs again?<- Can not happen!
And if let sync_dirty_inodes flush dentry page of inodeB, then inodeB
will sikp the
Since __write_data_page will not do f2fs_balance_fs for dir inode, so
there is no lock.
- f2fs_balance_fs
- __write_data_page (dir inode)
- f2fs_balance_fs again?<- Can not happen!
And if let sync_dirty_inodes flush dentry page of inodeB, then inodeB
will sikp the
Since __write_data_page will not do f2fs_balance_fs for dir inode, so
there is no lock.
- f2fs_balance_fs
- __write_data_page (dir inode)
- f2fs_balance_fs again?<- Can not happen!
And if let sync_dirty_inodes flush dentry page of inodeB, then inodeB
will sikp the
Since __write_data_page will not do f2fs_balance_fs for dir inode, so
there is no lock.
- f2fs_balance_fs
- __write_data_page (dir inode)
- f2fs_balance_fs again?<- Can not happen!
And if let sync_dirty_inodes flush dentry page of inodeB, then inodeB
will sikp the
On 08/02, Yunlong Song wrote:
> Hi Jay,
> The case is like this:
> write_checkpoint: normal dentry
> page writeback of inodeB:
> -block_operations -f2fs_write_data_pages
> -SBI_BLOCK_OPS is set -f2fs_write_cache_pages
>
On 08/02, Yunlong Song wrote:
> Hi Jay,
> The case is like this:
> write_checkpoint: normal dentry
> page writeback of inodeB:
> -block_operations -f2fs_write_data_pages
> -SBI_BLOCK_OPS is set -f2fs_write_cache_pages
>
Hi Jay,
The case is like this:
write_checkpoint: normal
dentry page writeback of inodeB:
-block_operations -f2fs_write_data_pages
-SBI_BLOCK_OPS is set -f2fs_write_cache_pages
-sync_dirty_inodes -__write_data_page
Hi Jay,
The case is like this:
write_checkpoint: normal
dentry page writeback of inodeB:
-block_operations -f2fs_write_data_pages
-SBI_BLOCK_OPS is set -f2fs_write_cache_pages
-sync_dirty_inodes -__write_data_page
On 08/01, Yunlong Song wrote:
> Hi Jay,
> The SBI_BLOCK_OPS can not cover all the case, once SBI_BLOCK_OPS is set,
> all the normal writeback
> (before the SBI_BLOCK_OPS is clear) of dentry pages which do not come from
> write_checkpoint flow will
> totally miss all the f2fs_balance_fs check.
On 08/01, Yunlong Song wrote:
> Hi Jay,
> The SBI_BLOCK_OPS can not cover all the case, once SBI_BLOCK_OPS is set,
> all the normal writeback
> (before the SBI_BLOCK_OPS is clear) of dentry pages which do not come from
> write_checkpoint flow will
> totally miss all the f2fs_balance_fs check.
Hi Jay,
The SBI_BLOCK_OPS can not cover all the case, once SBI_BLOCK_OPS is
set, all the normal writeback
(before the SBI_BLOCK_OPS is clear) of dentry pages which do not come
from write_checkpoint flow will
totally miss all the f2fs_balance_fs check.
On 2017/8/1 8:44, Jaegeuk Kim wrote:
Hi Jay,
The SBI_BLOCK_OPS can not cover all the case, once SBI_BLOCK_OPS is
set, all the normal writeback
(before the SBI_BLOCK_OPS is clear) of dentry pages which do not come
from write_checkpoint flow will
totally miss all the f2fs_balance_fs check.
On 2017/8/1 8:44, Jaegeuk Kim wrote:
Hi Yunlong,
On 07/29, Yunlong Song wrote:
> f2fs_balance_fs of dentry pages is skipped in __write_data_page due to
> deadlock
> of gc_mutex in write_checkpoint flow. This patch enables f2fs_balance_fs for
> normal dentry page writeback to ensure there are always enough free segments.
So, how
Hi Yunlong,
On 07/29, Yunlong Song wrote:
> f2fs_balance_fs of dentry pages is skipped in __write_data_page due to
> deadlock
> of gc_mutex in write_checkpoint flow. This patch enables f2fs_balance_fs for
> normal dentry page writeback to ensure there are always enough free segments.
So, how
Hi Jaegeuk,
On 2017/7/30 15:31, Jaegeuk Kim wrote:
> On 07/29, Yunlong Song wrote:
>> f2fs_balance_fs of dentry pages is skipped in __write_data_page due to
>> deadlock
>> of gc_mutex in write_checkpoint flow. This patch enables f2fs_balance_fs for
>> normal dentry page writeback to ensure there
Hi Jaegeuk,
On 2017/7/30 15:31, Jaegeuk Kim wrote:
> On 07/29, Yunlong Song wrote:
>> f2fs_balance_fs of dentry pages is skipped in __write_data_page due to
>> deadlock
>> of gc_mutex in write_checkpoint flow. This patch enables f2fs_balance_fs for
>> normal dentry page writeback to ensure there
On 07/29, Yunlong Song wrote:
> f2fs_balance_fs of dentry pages is skipped in __write_data_page due to
> deadlock
> of gc_mutex in write_checkpoint flow. This patch enables f2fs_balance_fs for
> normal dentry page writeback to ensure there are always enough free segments.
Sorry, by the way, why
On 07/29, Yunlong Song wrote:
> f2fs_balance_fs of dentry pages is skipped in __write_data_page due to
> deadlock
> of gc_mutex in write_checkpoint flow. This patch enables f2fs_balance_fs for
> normal dentry page writeback to ensure there are always enough free segments.
Sorry, by the way, why
Hi Yunlong,
[auto build test ERROR on f2fs/dev-test]
[also build test ERROR on v4.13-rc2 next-20170728]
[if your patch is applied to the wrong git tree, please drop us a note to help
improve the system]
url:
Hi Yunlong,
[auto build test ERROR on f2fs/dev-test]
[also build test ERROR on v4.13-rc2 next-20170728]
[if your patch is applied to the wrong git tree, please drop us a note to help
improve the system]
url:
f2fs_balance_fs of dentry pages is skipped in __write_data_page due to deadlock
of gc_mutex in write_checkpoint flow. This patch enables f2fs_balance_fs for
normal dentry page writeback to ensure there are always enough free segments.
Reported-by: Chao Yu
Signed-off-by:
f2fs_balance_fs of dentry pages is skipped in __write_data_page due to deadlock
of gc_mutex in write_checkpoint flow. This patch enables f2fs_balance_fs for
normal dentry page writeback to ensure there are always enough free segments.
Reported-by: Chao Yu
Signed-off-by: Yunlong Song
---
26 matches
Mail list logo