NFS/localio: prevent direct reclaim recursion into NFS via nfs_writepages

LOCALIO is an NFS loopback mount optimization that avoids using the
network for READ, WRITE and COMMIT if the NFS client and server are
determined to be on the same system. But because LOCALIO is still
fundamentally "just NFS loopback mount" it is susceptible to recursion
deadlock via direct reclaim, e.g.: NFS LOCALIO down to XFS and then
back into NFS via nfs_writepages.

Fix LOCALIO's potential for direct reclaim deadlock by ensuring that
all its page cache allocations are done from GFP_NOFS context.

Thanks to Ben Coddington for pointing out commit ad22c7a043c2 ("xfs:
prevent stack overflows from page cache allocation").

Reported-by: John Cagle <john.cagle@hammerspace.com>
Tested-by: Allen Lu <allen.lu@hammerspace.com>
Suggested-by: Benjamin Coddington <bcodding@hammerspace.com>
Fixes: 70ba381e1a43 ("nfs: add LOCALIO support")
Signed-off-by: Mike Snitzer <snitzer@hammerspace.com>
Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>

diff --git a/fs/nfs/localio.c b/fs/nfs/localio.c
index 7f5e7f0e3e00d5301e363fca5b22e683696ce5d8..40e20b324b3fc41242acede1b0b836334ba9b422 100644 (file)
--- a/fs/nfs/localio.c
+++ b/fs/nfs/localio.c
@@ -291,6 +291,18 @@ nfs_local_open_fh(struct nfs_client *clp, const struct cred *cred,
 }
 EXPORT_SYMBOL_GPL(nfs_local_open_fh);
 
+/*
+ * Ensure all page cache allocations are done from GFP_NOFS context to
+ * prevent direct reclaim recursion back into NFS via nfs_writepages.
+ */
+static void
+nfs_local_mapping_set_gfp_nofs_context(struct address_space *m)
+{
+       gfp_t gfp_mask = mapping_gfp_mask(m);
+
+       mapping_set_gfp_mask(m, (gfp_mask & ~(__GFP_FS)));
+}
+
 static void
 nfs_local_iocb_free(struct nfs_local_kiocb *iocb)
 {
@@ -315,6 +327,7 @@ nfs_local_iocb_alloc(struct nfs_pgio_header *hdr,
                return NULL;
        }
 
+       nfs_local_mapping_set_gfp_nofs_context(file->f_mapping);
        init_sync_kiocb(&iocb->kiocb, file);
 
        iocb->hdr = hdr;
@@ -1000,6 +1013,8 @@ nfs_local_run_commit(struct file *filp, struct nfs_commit_data *data)
                        end = LLONG_MAX;
        }
 
+       nfs_local_mapping_set_gfp_nofs_context(filp->f_mapping);
+
        dprintk("%s: commit %llu - %llu\n", __func__, start, end);
        return vfs_fsync_range(filp, start, end, 0);
 }