NFS: Always clear an invalid mapping when attempting a buffered write

If the page cache is invalid, then we can't do read-modify-write, so
ensure that we do clear it when we know it is invalid.

Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
This commit is contained in:
Trond Myklebust 2021-02-08 08:55:47 -05:00 committed by Anna Schumaker
parent fc9dc40189
commit 28aa2f9e73
3 changed files with 65 additions and 51 deletions

View File

@ -632,6 +632,8 @@ ssize_t nfs_file_write(struct kiocb *iocb, struct iov_iter *from)
goto out;
}
nfs_clear_invalid_mapping(file->f_mapping);
since = filemap_sample_wb_err(file->f_mapping);
nfs_start_io_write(inode);
result = generic_write_checks(iocb, from);

View File

@ -1257,6 +1257,63 @@ static int nfs_invalidate_mapping(struct inode *inode, struct address_space *map
return 0;
}
/**
* nfs_clear_invalid_mapping - Conditionally clear a mapping
* @mapping: pointer to mapping
*
* If the NFS_INO_INVALID_DATA inode flag is set, clear the mapping.
*/
int nfs_clear_invalid_mapping(struct address_space *mapping)
{
struct inode *inode = mapping->host;
struct nfs_inode *nfsi = NFS_I(inode);
unsigned long *bitlock = &nfsi->flags;
int ret = 0;
/*
* We must clear NFS_INO_INVALID_DATA first to ensure that
* invalidations that come in while we're shooting down the mappings
* are respected. But, that leaves a race window where one revalidator
* can clear the flag, and then another checks it before the mapping
* gets invalidated. Fix that by serializing access to this part of
* the function.
*
* At the same time, we need to allow other tasks to see whether we
* might be in the middle of invalidating the pages, so we only set
* the bit lock here if it looks like we're going to be doing that.
*/
for (;;) {
ret = wait_on_bit_action(bitlock, NFS_INO_INVALIDATING,
nfs_wait_bit_killable, TASK_KILLABLE);
if (ret)
goto out;
spin_lock(&inode->i_lock);
if (test_bit(NFS_INO_INVALIDATING, bitlock)) {
spin_unlock(&inode->i_lock);
continue;
}
if (nfsi->cache_validity & NFS_INO_INVALID_DATA)
break;
spin_unlock(&inode->i_lock);
goto out;
}
set_bit(NFS_INO_INVALIDATING, bitlock);
smp_wmb();
nfsi->cache_validity &=
~(NFS_INO_INVALID_DATA | NFS_INO_DATA_INVAL_DEFER);
spin_unlock(&inode->i_lock);
trace_nfs_invalidate_mapping_enter(inode);
ret = nfs_invalidate_mapping(inode, mapping);
trace_nfs_invalidate_mapping_exit(inode, ret);
clear_bit_unlock(NFS_INO_INVALIDATING, bitlock);
smp_mb__after_atomic();
wake_up_bit(bitlock, NFS_INO_INVALIDATING);
out:
return ret;
}
bool nfs_mapping_need_revalidate_inode(struct inode *inode)
{
return nfs_check_cache_invalid(inode, NFS_INO_REVAL_PAGECACHE) ||
@ -1289,67 +1346,21 @@ out:
* @inode: pointer to host inode
* @mapping: pointer to mapping
*/
int nfs_revalidate_mapping(struct inode *inode,
struct address_space *mapping)
int nfs_revalidate_mapping(struct inode *inode, struct address_space *mapping)
{
struct nfs_inode *nfsi = NFS_I(inode);
unsigned long *bitlock = &nfsi->flags;
int ret = 0;
/* swapfiles are not supposed to be shared. */
if (IS_SWAPFILE(inode))
goto out;
return 0;
if (nfs_mapping_need_revalidate_inode(inode)) {
ret = __nfs_revalidate_inode(NFS_SERVER(inode), inode);
int ret = __nfs_revalidate_inode(NFS_SERVER(inode), inode);
if (ret < 0)
goto out;
}
/*
* We must clear NFS_INO_INVALID_DATA first to ensure that
* invalidations that come in while we're shooting down the mappings
* are respected. But, that leaves a race window where one revalidator
* can clear the flag, and then another checks it before the mapping
* gets invalidated. Fix that by serializing access to this part of
* the function.
*
* At the same time, we need to allow other tasks to see whether we
* might be in the middle of invalidating the pages, so we only set
* the bit lock here if it looks like we're going to be doing that.
*/
for (;;) {
ret = wait_on_bit_action(bitlock, NFS_INO_INVALIDATING,
nfs_wait_bit_killable, TASK_KILLABLE);
if (ret)
goto out;
spin_lock(&inode->i_lock);
if (test_bit(NFS_INO_INVALIDATING, bitlock)) {
spin_unlock(&inode->i_lock);
continue;
}
if (nfsi->cache_validity & NFS_INO_INVALID_DATA)
break;
spin_unlock(&inode->i_lock);
goto out;
}
set_bit(NFS_INO_INVALIDATING, bitlock);
smp_wmb();
nfsi->cache_validity &= ~(NFS_INO_INVALID_DATA|
NFS_INO_DATA_INVAL_DEFER);
spin_unlock(&inode->i_lock);
trace_nfs_invalidate_mapping_enter(inode);
ret = nfs_invalidate_mapping(inode, mapping);
trace_nfs_invalidate_mapping_exit(inode, ret);
clear_bit_unlock(NFS_INO_INVALIDATING, bitlock);
smp_mb__after_atomic();
wake_up_bit(bitlock, NFS_INO_INVALIDATING);
out:
return ret;
}
return nfs_clear_invalid_mapping(mapping);
}
static bool nfs_file_has_writers(struct nfs_inode *nfsi)
{
struct inode *inode = &nfsi->vfs_inode;

View File

@ -387,6 +387,7 @@ extern int nfs_open(struct inode *, struct file *);
extern int nfs_attribute_cache_expired(struct inode *inode);
extern int nfs_revalidate_inode(struct nfs_server *server, struct inode *inode);
extern int __nfs_revalidate_inode(struct nfs_server *, struct inode *);
extern int nfs_clear_invalid_mapping(struct address_space *mapping);
extern bool nfs_mapping_need_revalidate_inode(struct inode *inode);
extern int nfs_revalidate_mapping(struct inode *inode, struct address_space *mapping);
extern int nfs_revalidate_mapping_rcu(struct inode *inode);