Hi Linux,
Could you please consider the below pull request?
Thanks,
The following changes since commit ace52288edf0cb5e7a52b681f057f86224c49c27:
Merge tag 'for-linus-20171218' of git://git.infradead.org/linux-mtd (2017-12-18 16:18:01 -0800)
are available in the Git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs.git tags/f2fs-for-4.16-rc1
for you to fetch changes up to 1c1d35df71104c76a4a2e25862926f22c334c9d6:
f2fs: support inode creation time (2018-01-25 14:10:39 -0800)
----------------------------------------------------------------
f2fs-for-4.16-rc1
In this round, we've followed up to support some generic features such as
cgroup, block reservation, linking fscrypt_ops, delivering write_hints,
and some ioctls. And, we could fix some corner cases in terms of power-cut
recovery and subtle deadlocks.
Enhancement:
- bitmap operations to handle NAT blocks
- readahead to improve readdir speed
- switch to use fscrypt_*
- apply write hints for direct IO
- add reserve_root=%u,resuid=%u,resgid=%u to reserve blocks for root/uid/gid
- modify b_avail and b_free to consider root reserved blocks
- support cgroup writeback
- support FIEMAP_FLAG_XATTR for fibmap
- add F2FS_IOC_PRECACHE_EXTENTS to pre-cache extents
- add F2FS_IOC_{GET/SET}_PIN_FILE to pin LBAs for data blocks
- support inode creation time
Bug fix:
- sysfile-based quota operations
- memory footprint accounting
- allow to write data on partial preallocation case
- fix deadlock case on fallocate
- fix to handle fill_super errors
- fix missing inode updates of fsync'ed file
- recover renamed file which was fsycn'ed before
- drop inmemory pages in corner error case
- keep last_disk_size correctly
- recover missing i_inline flags during roll-forward
Various clean-up patches were added as well.
----------------------------------------------------------------
Chao Yu (32):
f2fs: reserve nid resource for quota sysfile
f2fs: remove unneeded memory footprint accounting
f2fs: fix lock dependency in between dio_rwsem & i_mmap_sem
f2fs: inject fault to kzalloc
f2fs: inject fault to kvmalloc
f2fs: spread f2fs_k{m,z}alloc
f2fs: fix error handling in fill_super
f2fs: clean up hash codes
f2fs: clean up f2fs_map_blocks
f2fs: don't return value in truncate_data_blocks_range
f2fs: fix potential hangtask in f2fs_trace_pid
f2fs: continue to do direct IO if we only preallocate partial blocks
f2fs: clean up unneeded declaration
f2fs: avoid high cpu usage in discard thread
f2fs: remove unused pend_list_tag
f2fs: fix to caclulate required free section correctly
f2fs: fix to cover f2fs_inline_data_fiemap with inode_lock
f2fs: support FIEMAP_FLAG_XATTR
f2fs: support F2FS_IOC_PRECACHE_EXTENTS
f2fs: speed up defragment on sparse file
f2fs: fix to drop all inmem pages correctly
f2fs: clean up error path of fill_super
f2fs: kill F2FS_INLINE_XATTR_ADDRS for cleanup
f2fs: fix to update last_disk_size correctly
f2fs: split need_inplace_update
f2fs: hanlde error case in f2fs_ioc_shutdown
f2fs: stop gc/discard thread after fs shutdown
f2fs: drop page cache after fs shutdown
f2fs: use GFP_F2FS_ZERO for cleanup
f2fs: clean up duplicated assignment in init_discard_policy
f2fs: stop issuing discard if fs is readonly
f2fs: support inode creation time
Daeho Jeong (3):
f2fs: handle newly created page when revoking inmem pages
f2fs: prevent newly created inode from being dirtied incorrectly
f2fs: correct removexattr behavior for null valued extended attribute
Elena Reshetova (1):
posix_acl: convert posix_acl.a_refcount from atomic_t to refcount_t
Eric Biggers (5):
f2fs: switch to fscrypt_file_open()
f2fs: switch to fscrypt_prepare_link()
f2fs: switch to fscrypt_prepare_rename()
f2fs: switch to fscrypt_prepare_lookup()
f2fs: switch to fscrypt_prepare_setattr()
Hyunchul Lee (1):
f2fs: apply write hints to select the type of segment for direct write
Jaegeuk Kim (12):
f2fs: return error during fill_super
f2fs: recover directory operations by fsync
f2fs: fix missing error number for xattr operation
f2fs: skip stop_checkpoint for user data writes
f2fs: enable quota at remount from r to w
f2fs: show precise # of blocks that user/root can use
f2fs: add reserved blocks for root user
f2fs: add resgid and resuid to reserve root blocks
f2fs: add an ioctl to disable GC for specific file
f2fs: allow quota to use reserved blocks
f2fs: recover some i_inline flags
f2fs: allow to recover node blocks given updated checkpoint
LiFan (4):
f2fs: fix concurrent problem for updating free bitmap
f2fs: remove an excess variable
f2fs: remove a redundant conditional expression
f2fs: use unlikely for release case
Sheng Yong (4):
f2fs: introduce sysfs readdir_ra to readahead inode block in readdir
f2fs: still write data if preallocate only partial blocks
f2fs: remove unused parameter
f2fs: avoid hungtask when GC encrypted block if io_bits is set
Wei Yongjun (1):
f2fs: make local functions static
Yufen Yu (1):
f2fs: implement cgroup writeback support
Yunlei He (6):
f2fs: no need to read nat block if nat_block_bitmap is set
f2fs: no need return value in restore summary process
f2fs: fix an error case of missing update inode page
f2fs: update inode info to inode page for new file
f2fs: check node page again in write end io
f2fs: rebuild sit page from sit info in mem
Yunlong Song (1):
f2fs: check segment type in __f2fs_replace_block
Zhikang Zhang (1):
f2fs: remove repeated f2fs_bug_on
Documentation/ABI/testing/sysfs-fs-f2fs | 6 +
fs/f2fs/acl.c | 2 +-
fs/f2fs/checkpoint.c | 10 +-
fs/f2fs/data.c | 299 +++++++++++++++++++++++++++-----
fs/f2fs/debug.c | 12 +-
fs/f2fs/dir.c | 6 +
fs/f2fs/f2fs.h | 201 +++++++++++++++------
fs/f2fs/file.c | 248 ++++++++++++++++++++------
fs/f2fs/gc.c | 18 +-
fs/f2fs/gc.h | 2 +
fs/f2fs/inode.c | 34 +++-
fs/f2fs/namei.c | 67 +++----
fs/f2fs/node.c | 149 ++++++++--------
fs/f2fs/node.h | 4 +
fs/f2fs/recovery.c | 27 ++-
fs/f2fs/segment.c | 129 +++++++-------
fs/f2fs/segment.h | 92 +++++-----
fs/f2fs/super.c | 142 ++++++++++++---
fs/f2fs/sysfs.c | 14 +-
fs/f2fs/trace.c | 12 +-
fs/f2fs/xattr.c | 12 +-
fs/posix_acl.c | 6 +-
include/linux/f2fs_fs.h | 14 +-
include/linux/posix_acl.h | 7 +-
include/trace/events/f2fs.h | 3 +-
25 files changed, 1064 insertions(+), 452 deletions(-)
Hi all,
On Tue, 30 Jan 2018 14:32:20 -0800 Jaegeuk Kim <[email protected]> wrote:
>
> Could you please consider the below pull request?
>
> Thanks,
>
> The following changes since commit ace52288edf0cb5e7a52b681f057f86224c49c27:
>
> Merge tag 'for-linus-20171218' of git://git.infradead.org/linux-mtd (2017-12-18 16:18:01 -0800)
>
> are available in the Git repository at:
>
> git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs.git tags/f2fs-for-4.16-rc1
>
> for you to fetch changes up to 1c1d35df71104c76a4a2e25862926f22c334c9d6:
>
> f2fs: support inode creation time (2018-01-25 14:10:39 -0800)
This has an interesting conflict with the btrfs tree:
-----------------------------------------------------------------------
After merging the f2fs tree, today's linux-next build (x86_64
allmodconfig) failed like this:
/home/sfr/next/next/fs/f2fs/super.c: In function 'f2fs_fill_super':
/home/sfr/next/next/fs/f2fs/super.c:2563:18: error: 'SB_I_CGROUPWB' undeclared (first use in this function); did you mean 'S_CGROUPWB'?
sb->s_iflags |= SB_I_CGROUPWB;
^~~~~~~~~~~~~
S_CGROUPWB
Caused by commit
578c647879f7 ("f2fs: implement cgroup writeback support")
interacting with commit
34934189a3dd ("cgroup, writeback: replace SB_I_CGROUPWB with per-inode S_CGROUPWB"
from the btrfs tree.
I have reverted the f2fs tree commit for today, but a better resolution
would be appreciated.
-----------------------------------------------------------------------
I have seen no response to the above email and have been reverting that
commit since Jan 18.
--
Cheers,
Stephen Rothwell
On 01/31, Stephen Rothwell wrote:
> Hi all,
>
> On Tue, 30 Jan 2018 14:32:20 -0800 Jaegeuk Kim <[email protected]> wrote:
> >
> > Could you please consider the below pull request?
> >
> > Thanks,
> >
> > The following changes since commit ace52288edf0cb5e7a52b681f057f86224c49c27:
> >
> > Merge tag 'for-linus-20171218' of git://git.infradead.org/linux-mtd (2017-12-18 16:18:01 -0800)
> >
> > are available in the Git repository at:
> >
> > git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs.git tags/f2fs-for-4.16-rc1
> >
> > for you to fetch changes up to 1c1d35df71104c76a4a2e25862926f22c334c9d6:
> >
> > f2fs: support inode creation time (2018-01-25 14:10:39 -0800)
>
> This has an interesting conflict with the btrfs tree:
>
> -----------------------------------------------------------------------
> After merging the f2fs tree, today's linux-next build (x86_64
> allmodconfig) failed like this:
>
> /home/sfr/next/next/fs/f2fs/super.c: In function 'f2fs_fill_super':
> /home/sfr/next/next/fs/f2fs/super.c:2563:18: error: 'SB_I_CGROUPWB' undeclared (first use in this function); did you mean 'S_CGROUPWB'?
> sb->s_iflags |= SB_I_CGROUPWB;
> ^~~~~~~~~~~~~
> S_CGROUPWB
>
> Caused by commit
>
> 578c647879f7 ("f2fs: implement cgroup writeback support")
>
> interacting with commit
>
> 34934189a3dd ("cgroup, writeback: replace SB_I_CGROUPWB with per-inode S_CGROUPWB"
>
> from the btrfs tree.
>
> I have reverted the f2fs tree commit for today, but a better resolution
> would be appreciated.
> -----------------------------------------------------------------------
>
> I have seen no response to the above email and have been reverting that
> commit since Jan 18.
I think this can be addressed by:
diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
index 89c838bfb067..c743b73976d0 100644
--- a/fs/f2fs/inode.c
+++ b/fs/f2fs/inode.c
@@ -48,9 +48,10 @@ void f2fs_set_inode_flags(struct inode *inode)
new_fl |= S_DIRSYNC;
if (f2fs_encrypted_inode(inode))
new_fl |= S_ENCRYPTED;
+ new_fl |= S_CGROUPWB;
inode_set_flags(inode, new_fl,
S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC|
- S_ENCRYPTED);
+ S_ENCRYPTED|S_CGROUPWB);
}
static void __get_inode_rdev(struct inode *inode, struct f2fs_inode *ri)
diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
index 8173ae688814..ebcd58476fc7 100644
--- a/fs/f2fs/super.c
+++ b/fs/f2fs/super.c
@@ -2560,7 +2560,6 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent)
sb->s_flags = (sb->s_flags & ~SB_POSIXACL) |
(test_opt(sbi, POSIX_ACL) ? SB_POSIXACL : 0);
memcpy(&sb->s_uuid, raw_super->uuid, sizeof(raw_super->uuid));
- sb->s_iflags |= SB_I_CGROUPWB;
/* init f2fs-specific super block info */
sbi->valid_super_block = valid_super_block;
Hi Jaegeuk,
On Tue, 30 Jan 2018 16:25:35 -0800 Jaegeuk Kim <[email protected]> wrote:
>
> On 01/31, Stephen Rothwell wrote:
> >
> > On Tue, 30 Jan 2018 14:32:20 -0800 Jaegeuk Kim <[email protected]> wrote:
> > >
> > > Could you please consider the below pull request?
> > >
> > > Thanks,
> > >
> > > The following changes since commit ace52288edf0cb5e7a52b681f057f86224c49c27:
> > >
> > > Merge tag 'for-linus-20171218' of git://git.infradead.org/linux-mtd (2017-12-18 16:18:01 -0800)
> > >
> > > are available in the Git repository at:
> > >
> > > git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs.git tags/f2fs-for-4.16-rc1
> > >
> > > for you to fetch changes up to 1c1d35df71104c76a4a2e25862926f22c334c9d6:
> > >
> > > f2fs: support inode creation time (2018-01-25 14:10:39 -0800)
> >
> > This has an interesting conflict with the btrfs tree:
> >
> > -----------------------------------------------------------------------
> > After merging the f2fs tree, today's linux-next build (x86_64
> > allmodconfig) failed like this:
> >
> > /home/sfr/next/next/fs/f2fs/super.c: In function 'f2fs_fill_super':
> > /home/sfr/next/next/fs/f2fs/super.c:2563:18: error: 'SB_I_CGROUPWB' undeclared (first use in this function); did you mean 'S_CGROUPWB'?
> > sb->s_iflags |= SB_I_CGROUPWB;
> > ^~~~~~~~~~~~~
> > S_CGROUPWB
> >
> > Caused by commit
> >
> > 578c647879f7 ("f2fs: implement cgroup writeback support")
> >
> > interacting with commit
> >
> > 34934189a3dd ("cgroup, writeback: replace SB_I_CGROUPWB with per-inode S_CGROUPWB"
> >
> > from the btrfs tree.
> >
> > I have reverted the f2fs tree commit for today, but a better resolution
> > would be appreciated.
> > -----------------------------------------------------------------------
> >
> > I have seen no response to the above email and have been reverting that
> > commit since Jan 18.
>
> I think this can be addressed by:
>
> diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
> index 89c838bfb067..c743b73976d0 100644
> --- a/fs/f2fs/inode.c
> +++ b/fs/f2fs/inode.c
> @@ -48,9 +48,10 @@ void f2fs_set_inode_flags(struct inode *inode)
> new_fl |= S_DIRSYNC;
> if (f2fs_encrypted_inode(inode))
> new_fl |= S_ENCRYPTED;
> + new_fl |= S_CGROUPWB;
> inode_set_flags(inode, new_fl,
> S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC|
> - S_ENCRYPTED);
> + S_ENCRYPTED|S_CGROUPWB);
> }
>
> static void __get_inode_rdev(struct inode *inode, struct f2fs_inode *ri)
> diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
> index 8173ae688814..ebcd58476fc7 100644
> --- a/fs/f2fs/super.c
> +++ b/fs/f2fs/super.c
> @@ -2560,7 +2560,6 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent)
> sb->s_flags = (sb->s_flags & ~SB_POSIXACL) |
> (test_opt(sbi, POSIX_ACL) ? SB_POSIXACL : 0);
> memcpy(&sb->s_uuid, raw_super->uuid, sizeof(raw_super->uuid));
> - sb->s_iflags |= SB_I_CGROUPWB;
>
> /* init f2fs-specific super block info */
> sbi->valid_super_block = valid_super_block;
>
I have applied that (by hand) as a merge fix patch for the merge of teh
btrfs tree for today. Thanks for the resolution.
--
Cheers,
Stephen Rothwell
Hi all,
On Thu, 1 Feb 2018 09:25:51 +1100 Stephen Rothwell <[email protected]> wrote:
>
> On Tue, 30 Jan 2018 16:25:35 -0800 Jaegeuk Kim <[email protected]> wrote:
> >
> > On 01/31, Stephen Rothwell wrote:
> > >
> > > On Tue, 30 Jan 2018 14:32:20 -0800 Jaegeuk Kim <[email protected]> wrote:
> > > >
> > > > Could you please consider the below pull request?
> > > >
> > > > Thanks,
> > > >
> > > > The following changes since commit ace52288edf0cb5e7a52b681f057f86224c49c27:
> > > >
> > > > Merge tag 'for-linus-20171218' of git://git.infradead.org/linux-mtd (2017-12-18 16:18:01 -0800)
> > > >
> > > > are available in the Git repository at:
> > > >
> > > > git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs.git tags/f2fs-for-4.16-rc1
> > > >
> > > > for you to fetch changes up to 1c1d35df71104c76a4a2e25862926f22c334c9d6:
> > > >
> > > > f2fs: support inode creation time (2018-01-25 14:10:39 -0800)
> > >
> > > This has an interesting conflict with the btrfs tree:
> > >
> > > -----------------------------------------------------------------------
> > > After merging the f2fs tree, today's linux-next build (x86_64
> > > allmodconfig) failed like this:
> > >
> > > /home/sfr/next/next/fs/f2fs/super.c: In function 'f2fs_fill_super':
> > > /home/sfr/next/next/fs/f2fs/super.c:2563:18: error: 'SB_I_CGROUPWB' undeclared (first use in this function); did you mean 'S_CGROUPWB'?
> > > sb->s_iflags |= SB_I_CGROUPWB;
> > > ^~~~~~~~~~~~~
> > > S_CGROUPWB
> > >
> > > Caused by commit
> > >
> > > 578c647879f7 ("f2fs: implement cgroup writeback support")
> > >
> > > interacting with commit
> > >
> > > 34934189a3dd ("cgroup, writeback: replace SB_I_CGROUPWB with per-inode S_CGROUPWB"
> > >
> > > from the btrfs tree.
> > >
> > > I have reverted the f2fs tree commit for today, but a better resolution
> > > would be appreciated.
> > > -----------------------------------------------------------------------
> > >
> > > I have seen no response to the above email and have been reverting that
> > > commit since Jan 18.
> >
> > I think this can be addressed by:
> >
> > diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
> > index 89c838bfb067..c743b73976d0 100644
> > --- a/fs/f2fs/inode.c
> > +++ b/fs/f2fs/inode.c
> > @@ -48,9 +48,10 @@ void f2fs_set_inode_flags(struct inode *inode)
> > new_fl |= S_DIRSYNC;
> > if (f2fs_encrypted_inode(inode))
> > new_fl |= S_ENCRYPTED;
> > + new_fl |= S_CGROUPWB;
> > inode_set_flags(inode, new_fl,
> > S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC|
> > - S_ENCRYPTED);
> > + S_ENCRYPTED|S_CGROUPWB);
> > }
> >
> > static void __get_inode_rdev(struct inode *inode, struct f2fs_inode *ri)
> > diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
> > index 8173ae688814..ebcd58476fc7 100644
> > --- a/fs/f2fs/super.c
> > +++ b/fs/f2fs/super.c
> > @@ -2560,7 +2560,6 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent)
> > sb->s_flags = (sb->s_flags & ~SB_POSIXACL) |
> > (test_opt(sbi, POSIX_ACL) ? SB_POSIXACL : 0);
> > memcpy(&sb->s_uuid, raw_super->uuid, sizeof(raw_super->uuid));
> > - sb->s_iflags |= SB_I_CGROUPWB;
> >
> > /* init f2fs-specific super block info */
> > sbi->valid_super_block = valid_super_block;
> >
>
> I have applied that (by hand) as a merge fix patch for the merge of teh
> btrfs tree for today. Thanks for the resolution.
I am still applying this to the merge of the btrfs tree every day ...
Commit
578c647879f7 ("f2fs: implement cgroup writeback support")
was merged into Linus' tree on Jan 31.
Here is the actual merge fix patch:
From: Stephen Rothwell <[email protected]>
Date: Thu, 1 Feb 2018 09:21:06 +1100
Subject: [PATCH] f2fs: fixup for cgroup/writeback change
Signed-off-by: Stephen Rothwell <[email protected]>
---
fs/f2fs/inode.c | 3 ++-
fs/f2fs/super.c | 1 -
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
index 205add3d0f3a..f34652c4288f 100644
--- a/fs/f2fs/inode.c
+++ b/fs/f2fs/inode.c
@@ -48,9 +48,10 @@ void f2fs_set_inode_flags(struct inode *inode)
new_fl |= S_DIRSYNC;
if (f2fs_encrypted_inode(inode))
new_fl |= S_ENCRYPTED;
+ new_fl |= S_CGROUPWB;
inode_set_flags(inode, new_fl,
S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC|
- S_ENCRYPTED);
+ S_ENCRYPTED|S_CGROUPWB);
}
static void __get_inode_rdev(struct inode *inode, struct f2fs_inode *ri)
diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
index 8173ae688814..ebcd58476fc7 100644
--- a/fs/f2fs/super.c
+++ b/fs/f2fs/super.c
@@ -2560,7 +2560,6 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent)
sb->s_flags = (sb->s_flags & ~SB_POSIXACL) |
(test_opt(sbi, POSIX_ACL) ? SB_POSIXACL : 0);
memcpy(&sb->s_uuid, raw_super->uuid, sizeof(raw_super->uuid));
- sb->s_iflags |= SB_I_CGROUPWB;
/* init f2fs-specific super block info */
sbi->valid_super_block = valid_super_block;
--
2.15.1
--
Cheers,
Stephen Rothwell
Hi all,
Ping?
On Wed, 14 Mar 2018 10:17:04 +1100 Stephen Rothwell <[email protected]> wrote:
>
> On Thu, 1 Feb 2018 09:25:51 +1100 Stephen Rothwell <[email protected]> wrote:
> >
> > On Tue, 30 Jan 2018 16:25:35 -0800 Jaegeuk Kim <[email protected]> wrote:
> > >
> > > On 01/31, Stephen Rothwell wrote:
> > > >
> > > > On Tue, 30 Jan 2018 14:32:20 -0800 Jaegeuk Kim <[email protected]> wrote:
> > > > >
> > > > > Could you please consider the below pull request?
> > > > >
> > > > > Thanks,
> > > > >
> > > > > The following changes since commit ace52288edf0cb5e7a52b681f057f86224c49c27:
> > > > >
> > > > > Merge tag 'for-linus-20171218' of git://git.infradead.org/linux-mtd (2017-12-18 16:18:01 -0800)
> > > > >
> > > > > are available in the Git repository at:
> > > > >
> > > > > git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs.git tags/f2fs-for-4.16-rc1
> > > > >
> > > > > for you to fetch changes up to 1c1d35df71104c76a4a2e25862926f22c334c9d6:
> > > > >
> > > > > f2fs: support inode creation time (2018-01-25 14:10:39 -0800)
> > > >
> > > > This has an interesting conflict with the btrfs tree:
> > > >
> > > > -----------------------------------------------------------------------
> > > > After merging the f2fs tree, today's linux-next build (x86_64
> > > > allmodconfig) failed like this:
> > > >
> > > > /home/sfr/next/next/fs/f2fs/super.c: In function 'f2fs_fill_super':
> > > > /home/sfr/next/next/fs/f2fs/super.c:2563:18: error: 'SB_I_CGROUPWB' undeclared (first use in this function); did you mean 'S_CGROUPWB'?
> > > > sb->s_iflags |= SB_I_CGROUPWB;
> > > > ^~~~~~~~~~~~~
> > > > S_CGROUPWB
> > > >
> > > > Caused by commit
> > > >
> > > > 578c647879f7 ("f2fs: implement cgroup writeback support")
> > > >
> > > > interacting with commit
> > > >
> > > > 34934189a3dd ("cgroup, writeback: replace SB_I_CGROUPWB with per-inode S_CGROUPWB"
> > > >
> > > > from the btrfs tree.
> > > >
> > > > I have reverted the f2fs tree commit for today, but a better resolution
> > > > would be appreciated.
> > > > -----------------------------------------------------------------------
> > > >
> > > > I have seen no response to the above email and have been reverting that
> > > > commit since Jan 18.
> > >
> > > I think this can be addressed by:
> > >
> > > diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
> > > index 89c838bfb067..c743b73976d0 100644
> > > --- a/fs/f2fs/inode.c
> > > +++ b/fs/f2fs/inode.c
> > > @@ -48,9 +48,10 @@ void f2fs_set_inode_flags(struct inode *inode)
> > > new_fl |= S_DIRSYNC;
> > > if (f2fs_encrypted_inode(inode))
> > > new_fl |= S_ENCRYPTED;
> > > + new_fl |= S_CGROUPWB;
> > > inode_set_flags(inode, new_fl,
> > > S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC|
> > > - S_ENCRYPTED);
> > > + S_ENCRYPTED|S_CGROUPWB);
> > > }
> > >
> > > static void __get_inode_rdev(struct inode *inode, struct f2fs_inode *ri)
> > > diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
> > > index 8173ae688814..ebcd58476fc7 100644
> > > --- a/fs/f2fs/super.c
> > > +++ b/fs/f2fs/super.c
> > > @@ -2560,7 +2560,6 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent)
> > > sb->s_flags = (sb->s_flags & ~SB_POSIXACL) |
> > > (test_opt(sbi, POSIX_ACL) ? SB_POSIXACL : 0);
> > > memcpy(&sb->s_uuid, raw_super->uuid, sizeof(raw_super->uuid));
> > > - sb->s_iflags |= SB_I_CGROUPWB;
> > >
> > > /* init f2fs-specific super block info */
> > > sbi->valid_super_block = valid_super_block;
> > >
> >
> > I have applied that (by hand) as a merge fix patch for the merge of teh
> > btrfs tree for today. Thanks for the resolution.
>
> I am still applying this to the merge of the btrfs tree every day ...
>
> Commit
> 578c647879f7 ("f2fs: implement cgroup writeback support")
> was merged into Linus' tree on Jan 31.
>
> Here is the actual merge fix patch:
>
> From: Stephen Rothwell <[email protected]>
> Date: Thu, 1 Feb 2018 09:21:06 +1100
> Subject: [PATCH] f2fs: fixup for cgroup/writeback change
>
> Signed-off-by: Stephen Rothwell <[email protected]>
> ---
> fs/f2fs/inode.c | 3 ++-
> fs/f2fs/super.c | 1 -
> 2 files changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
> index 205add3d0f3a..f34652c4288f 100644
> --- a/fs/f2fs/inode.c
> +++ b/fs/f2fs/inode.c
> @@ -48,9 +48,10 @@ void f2fs_set_inode_flags(struct inode *inode)
> new_fl |= S_DIRSYNC;
> if (f2fs_encrypted_inode(inode))
> new_fl |= S_ENCRYPTED;
> + new_fl |= S_CGROUPWB;
> inode_set_flags(inode, new_fl,
> S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC|
> - S_ENCRYPTED);
> + S_ENCRYPTED|S_CGROUPWB);
> }
>
> static void __get_inode_rdev(struct inode *inode, struct f2fs_inode *ri)
> diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
> index 8173ae688814..ebcd58476fc7 100644
> --- a/fs/f2fs/super.c
> +++ b/fs/f2fs/super.c
> @@ -2560,7 +2560,6 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent)
> sb->s_flags = (sb->s_flags & ~SB_POSIXACL) |
> (test_opt(sbi, POSIX_ACL) ? SB_POSIXACL : 0);
> memcpy(&sb->s_uuid, raw_super->uuid, sizeof(raw_super->uuid));
> - sb->s_iflags |= SB_I_CGROUPWB;
>
> /* init f2fs-specific super block info */
> sbi->valid_super_block = valid_super_block;
> --
> 2.15.1
--
Cheers,
Stephen Rothwell
Hello, Stephen.
On Tue, Apr 03, 2018 at 12:29:19PM +1000, Stephen Rothwell wrote:
> > I am still applying this to the merge of the btrfs tree every day ...
> >
> > Commit
> > 578c647879f7 ("f2fs: implement cgroup writeback support")
> > was merged into Linus' tree on Jan 31.
> >
> > Here is the actual merge fix patch:
Yeah, the merge fix looks correct to me. Depending on the merge
order, it can go either after the btrfs merge or the f2fs merge, I
guess. Stephen, beyond appyling the merge fix afterwards, is there
somethine else you are concerned about?
Thanks.
--
tejun
Hi Tejun,
On Tue, 3 Apr 2018 07:20:29 -0700 Tejun Heo <[email protected]> wrote:
>
> Hello, Stephen.
>
> On Tue, Apr 03, 2018 at 12:29:19PM +1000, Stephen Rothwell wrote:
> > > I am still applying this to the merge of the btrfs tree every day ...
> > >
> > > Commit
> > > 578c647879f7 ("f2fs: implement cgroup writeback support")
> > > was merged into Linus' tree on Jan 31.
> > >
> > > Here is the actual merge fix patch:
>
> Yeah, the merge fix looks correct to me. Depending on the merge
> order, it can go either after the btrfs merge or the f2fs merge, I
> guess. Stephen, beyond appyling the merge fix afterwards, is there
> somethine else you are concerned about?
The f2fs part of the conflict was merged into Linus' tree on Jan 31,
and I have been applying the conflict resolution patch for a whole
release cycle i.e. the btrfs tree part of the conflict did not get
merged during the last merge window, and has not been updated since
then - in fact the btrfs tree
(git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs.git#next)
has not been updated or merged into Linus' tree since last December.
:-(
--
Cheers,
Stephen Rothwell
Hi all,
On Tue, 3 Apr 2018 12:29:19 +1000 Stephen Rothwell <[email protected]> wrote:
>
> Ping?
Ping again ...
So has Chris gone missing? Or should I just remove the btrfs tree from linux-next?
> On Wed, 14 Mar 2018 10:17:04 +1100 Stephen Rothwell <[email protected]> wrote:
> >
> > On Thu, 1 Feb 2018 09:25:51 +1100 Stephen Rothwell <[email protected]> wrote:
> > >
> > > On Tue, 30 Jan 2018 16:25:35 -0800 Jaegeuk Kim <[email protected]> wrote:
> > > >
> > > > On 01/31, Stephen Rothwell wrote:
> > > > >
> > > > > On Tue, 30 Jan 2018 14:32:20 -0800 Jaegeuk Kim <[email protected]> wrote:
> > > > > >
> > > > > > Could you please consider the below pull request?
> > > > > >
> > > > > > Thanks,
> > > > > >
> > > > > > The following changes since commit ace52288edf0cb5e7a52b681f057f86224c49c27:
> > > > > >
> > > > > > Merge tag 'for-linus-20171218' of git://git.infradead.org/linux-mtd (2017-12-18 16:18:01 -0800)
> > > > > >
> > > > > > are available in the Git repository at:
> > > > > >
> > > > > > git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs.git tags/f2fs-for-4.16-rc1
> > > > > >
> > > > > > for you to fetch changes up to 1c1d35df71104c76a4a2e25862926f22c334c9d6:
> > > > > >
> > > > > > f2fs: support inode creation time (2018-01-25 14:10:39 -0800)
> > > > >
> > > > > This has an interesting conflict with the btrfs tree:
> > > > >
> > > > > -----------------------------------------------------------------------
> > > > > After merging the f2fs tree, today's linux-next build (x86_64
> > > > > allmodconfig) failed like this:
> > > > >
> > > > > /home/sfr/next/next/fs/f2fs/super.c: In function 'f2fs_fill_super':
> > > > > /home/sfr/next/next/fs/f2fs/super.c:2563:18: error: 'SB_I_CGROUPWB' undeclared (first use in this function); did you mean 'S_CGROUPWB'?
> > > > > sb->s_iflags |= SB_I_CGROUPWB;
> > > > > ^~~~~~~~~~~~~
> > > > > S_CGROUPWB
> > > > >
> > > > > Caused by commit
> > > > >
> > > > > 578c647879f7 ("f2fs: implement cgroup writeback support")
> > > > >
> > > > > interacting with commit
> > > > >
> > > > > 34934189a3dd ("cgroup, writeback: replace SB_I_CGROUPWB with per-inode S_CGROUPWB"
> > > > >
> > > > > from the btrfs tree.
> > > > >
> > > > > I have reverted the f2fs tree commit for today, but a better resolution
> > > > > would be appreciated.
> > > > > -----------------------------------------------------------------------
> > > > >
> > > > > I have seen no response to the above email and have been reverting that
> > > > > commit since Jan 18.
> > > >
> > > > I think this can be addressed by:
> > > >
> > > > diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
> > > > index 89c838bfb067..c743b73976d0 100644
> > > > --- a/fs/f2fs/inode.c
> > > > +++ b/fs/f2fs/inode.c
> > > > @@ -48,9 +48,10 @@ void f2fs_set_inode_flags(struct inode *inode)
> > > > new_fl |= S_DIRSYNC;
> > > > if (f2fs_encrypted_inode(inode))
> > > > new_fl |= S_ENCRYPTED;
> > > > + new_fl |= S_CGROUPWB;
> > > > inode_set_flags(inode, new_fl,
> > > > S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC|
> > > > - S_ENCRYPTED);
> > > > + S_ENCRYPTED|S_CGROUPWB);
> > > > }
> > > >
> > > > static void __get_inode_rdev(struct inode *inode, struct f2fs_inode *ri)
> > > > diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
> > > > index 8173ae688814..ebcd58476fc7 100644
> > > > --- a/fs/f2fs/super.c
> > > > +++ b/fs/f2fs/super.c
> > > > @@ -2560,7 +2560,6 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent)
> > > > sb->s_flags = (sb->s_flags & ~SB_POSIXACL) |
> > > > (test_opt(sbi, POSIX_ACL) ? SB_POSIXACL : 0);
> > > > memcpy(&sb->s_uuid, raw_super->uuid, sizeof(raw_super->uuid));
> > > > - sb->s_iflags |= SB_I_CGROUPWB;
> > > >
> > > > /* init f2fs-specific super block info */
> > > > sbi->valid_super_block = valid_super_block;
> > > >
> > >
> > > I have applied that (by hand) as a merge fix patch for the merge of teh
> > > btrfs tree for today. Thanks for the resolution.
> >
> > I am still applying this to the merge of the btrfs tree every day ...
> >
> > Commit
> > 578c647879f7 ("f2fs: implement cgroup writeback support")
> > was merged into Linus' tree on Jan 31.
> >
> > Here is the actual merge fix patch:
> >
> > From: Stephen Rothwell <[email protected]>
> > Date: Thu, 1 Feb 2018 09:21:06 +1100
> > Subject: [PATCH] f2fs: fixup for cgroup/writeback change
> >
> > Signed-off-by: Stephen Rothwell <[email protected]>
> > ---
> > fs/f2fs/inode.c | 3 ++-
> > fs/f2fs/super.c | 1 -
> > 2 files changed, 2 insertions(+), 2 deletions(-)
> >
> > diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
> > index 205add3d0f3a..f34652c4288f 100644
> > --- a/fs/f2fs/inode.c
> > +++ b/fs/f2fs/inode.c
> > @@ -48,9 +48,10 @@ void f2fs_set_inode_flags(struct inode *inode)
> > new_fl |= S_DIRSYNC;
> > if (f2fs_encrypted_inode(inode))
> > new_fl |= S_ENCRYPTED;
> > + new_fl |= S_CGROUPWB;
> > inode_set_flags(inode, new_fl,
> > S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC|
> > - S_ENCRYPTED);
> > + S_ENCRYPTED|S_CGROUPWB);
> > }
> >
> > static void __get_inode_rdev(struct inode *inode, struct f2fs_inode *ri)
> > diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
> > index 8173ae688814..ebcd58476fc7 100644
> > --- a/fs/f2fs/super.c
> > +++ b/fs/f2fs/super.c
> > @@ -2560,7 +2560,6 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent)
> > sb->s_flags = (sb->s_flags & ~SB_POSIXACL) |
> > (test_opt(sbi, POSIX_ACL) ? SB_POSIXACL : 0);
> > memcpy(&sb->s_uuid, raw_super->uuid, sizeof(raw_super->uuid));
> > - sb->s_iflags |= SB_I_CGROUPWB;
> >
> > /* init f2fs-specific super block info */
> > sbi->valid_super_block = valid_super_block;
> > --
> > 2.15.1
--
Cheers,
Stephen Rothwell
On Thu, May 03, 2018 at 08:42:59AM +1000, Stephen Rothwell wrote:
> Hi all,
>
> On Tue, 3 Apr 2018 12:29:19 +1000 Stephen Rothwell <[email protected]> wrote:
> >
> > Ping?
>
> Ping again ...
>
> So has Chris gone missing? Or should I just remove the btrfs tree from linux-next?
As far as btrfs code changes are concerned, they go through my tree. The
other conflicting changes are some cgroup tweaks that have impact on
btrfs but do not affect the development.
Hi Chris,
On Tue, 3 Apr 2018 12:29:19 +1000 Stephen Rothwell <[email protected]> wrote:
>
> Ping?
>
> On Wed, 14 Mar 2018 10:17:04 +1100 Stephen Rothwell <[email protected]> wrote:
> >
> > On Thu, 1 Feb 2018 09:25:51 +1100 Stephen Rothwell <[email protected]> wrote:
> > >
> > > On Tue, 30 Jan 2018 16:25:35 -0800 Jaegeuk Kim <[email protected]> wrote:
> > > >
> > > > On 01/31, Stephen Rothwell wrote:
> > > > >
> > > > > On Tue, 30 Jan 2018 14:32:20 -0800 Jaegeuk Kim <[email protected]> wrote:
> > > > > >
> > > > > > Could you please consider the below pull request?
> > > > > >
> > > > > > Thanks,
> > > > > >
> > > > > > The following changes since commit ace52288edf0cb5e7a52b681f057f86224c49c27:
> > > > > >
> > > > > > Merge tag 'for-linus-20171218' of git://git.infradead.org/linux-mtd (2017-12-18 16:18:01 -0800)
> > > > > >
> > > > > > are available in the Git repository at:
> > > > > >
> > > > > > git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs.git tags/f2fs-for-4.16-rc1
> > > > > >
> > > > > > for you to fetch changes up to 1c1d35df71104c76a4a2e25862926f22c334c9d6:
> > > > > >
> > > > > > f2fs: support inode creation time (2018-01-25 14:10:39 -0800)
> > > > >
> > > > > This has an interesting conflict with the btrfs tree:
> > > > >
> > > > > -----------------------------------------------------------------------
> > > > > After merging the f2fs tree, today's linux-next build (x86_64
> > > > > allmodconfig) failed like this:
> > > > >
> > > > > /home/sfr/next/next/fs/f2fs/super.c: In function 'f2fs_fill_super':
> > > > > /home/sfr/next/next/fs/f2fs/super.c:2563:18: error: 'SB_I_CGROUPWB' undeclared (first use in this function); did you mean 'S_CGROUPWB'?
> > > > > sb->s_iflags |= SB_I_CGROUPWB;
> > > > > ^~~~~~~~~~~~~
> > > > > S_CGROUPWB
> > > > >
> > > > > Caused by commit
> > > > >
> > > > > 578c647879f7 ("f2fs: implement cgroup writeback support")
> > > > >
> > > > > interacting with commit
> > > > >
> > > > > 34934189a3dd ("cgroup, writeback: replace SB_I_CGROUPWB with per-inode S_CGROUPWB"
> > > > >
> > > > > from the btrfs tree.
> > > > >
> > > > > I have reverted the f2fs tree commit for today, but a better resolution
> > > > > would be appreciated.
> > > > > -----------------------------------------------------------------------
> > > > >
> > > > > I have seen no response to the above email and have been reverting that
> > > > > commit since Jan 18.
> > > >
> > > > I think this can be addressed by:
> > > >
> > > > diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
> > > > index 89c838bfb067..c743b73976d0 100644
> > > > --- a/fs/f2fs/inode.c
> > > > +++ b/fs/f2fs/inode.c
> > > > @@ -48,9 +48,10 @@ void f2fs_set_inode_flags(struct inode *inode)
> > > > new_fl |= S_DIRSYNC;
> > > > if (f2fs_encrypted_inode(inode))
> > > > new_fl |= S_ENCRYPTED;
> > > > + new_fl |= S_CGROUPWB;
> > > > inode_set_flags(inode, new_fl,
> > > > S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC|
> > > > - S_ENCRYPTED);
> > > > + S_ENCRYPTED|S_CGROUPWB);
> > > > }
> > > >
> > > > static void __get_inode_rdev(struct inode *inode, struct f2fs_inode *ri)
> > > > diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
> > > > index 8173ae688814..ebcd58476fc7 100644
> > > > --- a/fs/f2fs/super.c
> > > > +++ b/fs/f2fs/super.c
> > > > @@ -2560,7 +2560,6 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent)
> > > > sb->s_flags = (sb->s_flags & ~SB_POSIXACL) |
> > > > (test_opt(sbi, POSIX_ACL) ? SB_POSIXACL : 0);
> > > > memcpy(&sb->s_uuid, raw_super->uuid, sizeof(raw_super->uuid));
> > > > - sb->s_iflags |= SB_I_CGROUPWB;
> > > >
> > > > /* init f2fs-specific super block info */
> > > > sbi->valid_super_block = valid_super_block;
> > > >
> > >
> > > I have applied that (by hand) as a merge fix patch for the merge of teh
> > > btrfs tree for today. Thanks for the resolution.
> >
> > I am still applying this to the merge of the btrfs tree every day ...
> >
> > Commit
> > 578c647879f7 ("f2fs: implement cgroup writeback support")
> > was merged into Linus' tree on Jan 31.
> >
> > Here is the actual merge fix patch:
> >
> > From: Stephen Rothwell <[email protected]>
> > Date: Thu, 1 Feb 2018 09:21:06 +1100
> > Subject: [PATCH] f2fs: fixup for cgroup/writeback change
> >
> > Signed-off-by: Stephen Rothwell <[email protected]>
> > ---
> > fs/f2fs/inode.c | 3 ++-
> > fs/f2fs/super.c | 1 -
> > 2 files changed, 2 insertions(+), 2 deletions(-)
> >
> > diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
> > index 205add3d0f3a..f34652c4288f 100644
> > --- a/fs/f2fs/inode.c
> > +++ b/fs/f2fs/inode.c
> > @@ -48,9 +48,10 @@ void f2fs_set_inode_flags(struct inode *inode)
> > new_fl |= S_DIRSYNC;
> > if (f2fs_encrypted_inode(inode))
> > new_fl |= S_ENCRYPTED;
> > + new_fl |= S_CGROUPWB;
> > inode_set_flags(inode, new_fl,
> > S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC|
> > - S_ENCRYPTED);
> > + S_ENCRYPTED|S_CGROUPWB);
> > }
> >
> > static void __get_inode_rdev(struct inode *inode, struct f2fs_inode *ri)
> > diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
> > index 8173ae688814..ebcd58476fc7 100644
> > --- a/fs/f2fs/super.c
> > +++ b/fs/f2fs/super.c
> > @@ -2560,7 +2560,6 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent)
> > sb->s_flags = (sb->s_flags & ~SB_POSIXACL) |
> > (test_opt(sbi, POSIX_ACL) ? SB_POSIXACL : 0);
> > memcpy(&sb->s_uuid, raw_super->uuid, sizeof(raw_super->uuid));
> > - sb->s_iflags |= SB_I_CGROUPWB;
> >
> > /* init f2fs-specific super block info */
> > sbi->valid_super_block = valid_super_block;
> > --
> > 2.15.1
Another merge window has passed and I am still holding this tree and its
merge fixes ... Tomorrow I will drop the btrfs tree and rename the
btrfs-kdave tree to btrfs. OK?
--
Cheers,
Stephen Rothwell
On 17 Jun 2018, at 23:38, Stephen Rothwell wrote:
> Hi Chris,
>
>
> Another merge window has passed and I am still holding this tree and
> its
> merge fixes ... Tomorrow I will drop the btrfs tree and rename the
> btrfs-kdave tree to btrfs. OK?
Really sorry for the hassle, the f2fs subject line got this filtered
into a different inbox. Tejun and Josef are working out these btrfs
cgroup fixes, so I've pushed v4.17 to my -next tree for now.
Dave has been doing the Btrfs merge window for some time, so focusing
linux-next on his tree is a better fit. But either way, I've fixed my
my email filters so you're back at the top where you belong.
-chris
Hi Chris,
On Mon, 18 Jun 2018 14:05:32 -0400 Chris Mason <[email protected]> wrote:
>
> Really sorry for the hassle, the f2fs subject line got this filtered
> into a different inbox. Tejun and Josef are working out these btrfs
Ah, that explains it.
> cgroup fixes, so I've pushed v4.17 to my -next tree for now.
Thanks.
> Dave has been doing the Btrfs merge window for some time, so focusing
> linux-next on his tree is a better fit. But either way, I've fixed my
> my email filters so you're back at the top where you belong.
I'll just leave things as they are for now.
--
Cheers,
Stephen Rothwell