2020-05-20 01:24:34

by Zheng Bin

[permalink] [raw]
Subject: [PATCH] nfs: set invalid blocks after NFSv4 writes

Use the following command to test nfsv4(size of file1M is 1MB):
mount -t nfs -o vers=4.0,actimeo=60 127.0.0.1/dir1 /mnt
cp file1M /mnt
du -h /mnt/file1M -->0 within 60s, then 1M

When write is done(cp file1M /mnt), will call this:
nfs_writeback_done
nfs4_write_done
nfs4_write_done_cb
nfs_writeback_update_inode
nfs_post_op_update_inode_force_wcc_locked(change, ctime, mtime
nfs_post_op_update_inode_force_wcc_locked
nfs_set_cache_invalid
nfs_refresh_inode_locked
nfs_update_inode

nfsd write response contains change, ctime, mtime, the flag will be
clear after nfs_update_inode. Howerver, write response does not contain
space_used, previous open response contains space_used whose value is 0,
so inode->i_blocks is still 0.

nfs_getattr -->called by "du -h"
do_update |= force_sync || nfs_attribute_cache_expired -->false in 60s
cache_validity = READ_ONCE(NFS_I(inode)->cache_validity)
do_update |= cache_validity & (NFS_INO_INVALID_ATTR -->false
if (do_update) {
__nfs_revalidate_inode
}

Within 60s, does not send getattr request to nfsd, thus "du -h /mnt/file1M"
is 0.

Add a NFS_INO_INVALID_BLOCKS flag, set it when nfsv4 write is done.

Fixes: 16e143751727 ("NFS: More fine grained attribute tracking")
Signed-off-by: Zheng Bin <[email protected]>
---
fs/nfs/inode.c | 9 +++++++--
include/linux/nfs_fs.h | 6 ++++--
2 files changed, 11 insertions(+), 4 deletions(-)

diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c
index b9d0921cb4fe..2d743e42fee1 100644
--- a/fs/nfs/inode.c
+++ b/fs/nfs/inode.c
@@ -1764,7 +1764,8 @@ int nfs_post_op_update_inode_force_wcc_locked(struct inode *inode, struct nfs_fa
status = nfs_post_op_update_inode_locked(inode, fattr,
NFS_INO_INVALID_CHANGE
| NFS_INO_INVALID_CTIME
- | NFS_INO_INVALID_MTIME);
+ | NFS_INO_INVALID_MTIME
+ | NFS_INO_INVALID_BLOCKS);
return status;
}

@@ -2033,8 +2034,12 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
inode->i_blocks = nfs_calc_block_size(fattr->du.nfs3.used);
} else if (fattr->valid & NFS_ATTR_FATTR_BLOCKS_USED)
inode->i_blocks = fattr->du.nfs2.blocks;
- else
+ else {
+ nfsi->cache_validity |= save_cache_validity &
+ (NFS_INO_INVALID_BLOCKS
+ | NFS_INO_REVAL_FORCED);
cache_revalidated = false;
+ }

/* Update attrtimeo value if we're out of the unstable period */
if (attr_changed) {
diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h
index 73eda45f1cfd..77efdca12cab 100644
--- a/include/linux/nfs_fs.h
+++ b/include/linux/nfs_fs.h
@@ -227,14 +227,16 @@ struct nfs4_copy_state {
#define NFS_INO_INVALID_CTIME BIT(9) /* cached ctime is invalid */
#define NFS_INO_INVALID_MTIME BIT(10) /* cached mtime is invalid */
#define NFS_INO_INVALID_SIZE BIT(11) /* cached size is invalid */
-#define NFS_INO_INVALID_OTHER BIT(12) /* other attrs are invalid */
+#define NFS_INO_INVALID_BLOCKS BIT(12) /* cached blocks are invalid */
+#define NFS_INO_INVALID_OTHER BIT(13) /* other attrs are invalid */
#define NFS_INO_DATA_INVAL_DEFER \
- BIT(13) /* Deferred cache invalidation */
+ BIT(14) /* Deferred cache invalidation */

#define NFS_INO_INVALID_ATTR (NFS_INO_INVALID_CHANGE \
| NFS_INO_INVALID_CTIME \
| NFS_INO_INVALID_MTIME \
| NFS_INO_INVALID_SIZE \
+ | NFS_INO_INVALID_BLOCKS \
| NFS_INO_INVALID_OTHER) /* inode metadata is invalid */

/*
--
2.26.0.106.g9fadedd


2020-05-20 01:55:13

by Trond Myklebust

[permalink] [raw]
Subject: Re: [PATCH] nfs: set invalid blocks after NFSv4 writes

On Wed, 2020-05-20 at 09:30 +0800, Zheng Bin wrote:
> Use the following command to test nfsv4(size of file1M is 1MB):
> mount -t nfs -o vers=4.0,actimeo=60 127.0.0.1/dir1 /mnt
> cp file1M /mnt
> du -h /mnt/file1M -->0 within 60s, then 1M
>
> When write is done(cp file1M /mnt), will call this:
> nfs_writeback_done
> nfs4_write_done
> nfs4_write_done_cb
> nfs_writeback_update_inode
> nfs_post_op_update_inode_force_wcc_locked(change, ctime,
> mtime
> nfs_post_op_update_inode_force_wcc_locked
> nfs_set_cache_invalid
> nfs_refresh_inode_locked
> nfs_update_inode
>
> nfsd write response contains change, ctime, mtime, the flag will be
> clear after nfs_update_inode. Howerver, write response does not
> contain
> space_used, previous open response contains space_used whose value is
> 0,
> so inode->i_blocks is still 0.
>
> nfs_getattr -->called by "du -h"
> do_update |= force_sync || nfs_attribute_cache_expired -->false in
> 60s
> cache_validity = READ_ONCE(NFS_I(inode)->cache_validity)
> do_update |= cache_validity & (NFS_INO_INVALID_ATTR -->false
> if (do_update) {
> __nfs_revalidate_inode
> }
>
> Within 60s, does not send getattr request to nfsd, thus "du -h
> /mnt/file1M"
> is 0.
>
> Add a NFS_INO_INVALID_BLOCKS flag, set it when nfsv4 write is done.
>
> Fixes: 16e143751727 ("NFS: More fine grained attribute tracking")
> Signed-off-by: Zheng Bin <[email protected]>
> ---
> fs/nfs/inode.c | 9 +++++++--
> include/linux/nfs_fs.h | 6 ++++--
> 2 files changed, 11 insertions(+), 4 deletions(-)
>
> diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c
> index b9d0921cb4fe..2d743e42fee1 100644
> --- a/fs/nfs/inode.c
> +++ b/fs/nfs/inode.c
> @@ -1764,7 +1764,8 @@ int
> nfs_post_op_update_inode_force_wcc_locked(struct inode *inode, struct
> nfs_fa
> status = nfs_post_op_update_inode_locked(inode, fattr,
> NFS_INO_INVALID_CHANGE
> | NFS_INO_INVALID_CTIME
> - | NFS_INO_INVALID_MTIME);
> + | NFS_INO_INVALID_MTIME
> + | NFS_INO_INVALID_BLOCKS);
> return status;
> }
>
> @@ -2033,8 +2034,12 @@ static int nfs_update_inode(struct inode
> *inode, struct nfs_fattr *fattr)
> inode->i_blocks = nfs_calc_block_size(fattr-
> >du.nfs3.used);
> } else if (fattr->valid & NFS_ATTR_FATTR_BLOCKS_USED)
> inode->i_blocks = fattr->du.nfs2.blocks;
> - else
> + else {
> + nfsi->cache_validity |= save_cache_validity &
> + (NFS_INO_INVALID_BLOCKS
> + | NFS_INO_REVAL_FORCED);
> cache_revalidated = false;
> + }
>
> /* Update attrtimeo value if we're out of the unstable period
> */
> if (attr_changed) {
> diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h
> index 73eda45f1cfd..77efdca12cab 100644
> --- a/include/linux/nfs_fs.h
> +++ b/include/linux/nfs_fs.h
> @@ -227,14 +227,16 @@ struct nfs4_copy_state {
> #define NFS_INO_INVALID_CTIME BIT(9) /* cached
> ctime is invalid */
> #define NFS_INO_INVALID_MTIME BIT(10) /* cached
> mtime is invalid */
> #define NFS_INO_INVALID_SIZE BIT(11) /* cached size is
> invalid */
> -#define NFS_INO_INVALID_OTHER BIT(12) /* other
> attrs are invalid */
> +#define NFS_INO_INVALID_BLOCKS BIT(12) /* cached
> blocks are invalid */
> +#define NFS_INO_INVALID_OTHER BIT(13) /* other
> attrs are invalid */
> #define NFS_INO_DATA_INVAL_DEFER \
> - BIT(13) /* Deferred cache
> invalidation */
> + BIT(14) /* Deferred cache
> invalidation */
>
> #define NFS_INO_INVALID_ATTR (NFS_INO_INVALID_CHANGE \
> | NFS_INO_INVALID_CTIME \
> | NFS_INO_INVALID_MTIME \
> | NFS_INO_INVALID_SIZE \
> + | NFS_INO_INVALID_BLOCKS \

Can we please replace this with an explicit check for STATX_BLOCKS in
nfs_getattr()? The above would just mean we always revalidate
independently of whether or not the user cares.

> | NFS_INO_INVALID_OTHER) /* inode metadata is invalid
> */
>
> /*
> --
> 2.26.0.106.g9fadedd
>
--
Trond Myklebust
Linux NFS client maintainer, Hammerspace
[email protected]