pnfs/blocklayout: reject pnfs blocksize larger than page size
authorChristoph Hellwig <hch@lst.de>
Thu, 21 Aug 2014 16:09:26 +0000 (11:09 -0500)
committerTrond Myklebust <trond.myklebust@primarydata.com>
Wed, 10 Sep 2014 19:47:02 +0000 (12:47 -0700)
The Linux VM subsystem can't support block sizes larger than page size
for block based filesystems very well.  While this can be hacked around
to some extent for simple filesystems the read-modify-write cycles
required for pnfs block invalid extents are extremly deadlock prone
when operating on multiple pages.  Reject this case early on instead
of pretending to support it (badly).

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
fs/nfs/blocklayout/blocklayout.c

index cbb1797149d5731a77f22b97fc4a3ba2cdfbacfd..6c1a4212919dc2272608655aa6ee48a6089d7373 100644 (file)
@@ -1115,6 +1115,12 @@ bl_set_layoutdriver(struct nfs_server *server, const struct nfs_fh *fh)
                dprintk("%s Server did not return blksize\n", __func__);
                return -EINVAL;
        }
+       if (server->pnfs_blksize > PAGE_SIZE) {
+               printk(KERN_ERR "%s: pNFS blksize %d not supported.\n",
+                       __func__, server->pnfs_blksize);
+               return -EINVAL;
+       }
+
        b_mt_id = kzalloc(sizeof(struct block_mount_id), GFP_NOFS);
        if (!b_mt_id) {
                status = -ENOMEM;