tmakin/ack-tegra
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

NFS: O_DIRECT writes must check and adjust the file length

Browse Source 
[ Upstream commit fcf857ee1958e9247298251f7615d0c76f1e9b38 ]

While it is uncommon for delegations to be held while O_DIRECT writes
are in progress, it is possible. The xfstests generic/647 and
generic/729 both end up triggering that state, and end up failing due to
the fact that the file size is not adjusted.

Reported-by: Chuck Lever <chuck.lever@oracle.com>
Link: https://bugzilla.kernel.org/show_bug.cgi?id=219738
Cc: stable@vger.kernel.org
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
Stable-dep-of: 88025c67fe3c ("NFS: Adjust delegated timestamps for O_DIRECT reads and writes")
Signed-off-by: Sasha Levin <sashal@kernel.org>
This commit is contained in:
Trond Myklebust
2025-02-01 14:59:02 -05:00
committed by Greg Kroah-Hartman
parent 3594aad97e
commit b7b3d772e4
1 changed files with 19 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
fs/nfs/direct.c
+19
View File
@@ -130,6 +130,20 @@ static void nfs_direct_truncate_request(struct nfs_direct_req *dreq,
dreq->count = req_start;
}
static void nfs_direct_file_adjust_size_locked(struct inode *inode,
loff_t offset, size_t count)
{
loff_t newsize = offset + (loff_t)count;
loff_t oldsize = i_size_read(inode);
if (newsize > oldsize) {
i_size_write(inode, newsize);
NFS_I(inode)->cache_validity &= ~NFS_INO_INVALID_SIZE;
trace_nfs_size_grow(inode, newsize);
nfs_inc_stats(inode, NFSIOS_EXTENDWRITE);
}
}
/**
* nfs_swap_rw - NFS address space operation for swap I/O
* @iocb: target I/O control block
@@ -732,6 +746,7 @@ static void nfs_direct_write_completion(struct nfs_pgio_header *hdr)
struct nfs_direct_req *dreq = hdr->dreq;
struct nfs_commit_info cinfo;
struct nfs_page *req = nfs_list_entry(hdr->pages.next);
struct inode *inode = dreq->inode;
int flags = NFS_ODIRECT_DONE;
trace_nfs_direct_write_completion(dreq);
@@ -753,6 +768,10 @@ static void nfs_direct_write_completion(struct nfs_pgio_header *hdr)
}
spin_unlock(&dreq->lock);
spin_lock(&inode->i_lock);
nfs_direct_file_adjust_size_locked(inode, dreq->io_start, dreq->count);
spin_unlock(&inode->i_lock);
while (!list_empty(&hdr->pages)) {
req = nfs_list_entry(hdr->pages.next);