NFS: Don't use page_file_mapping after removing the page
authorBenjamin Coddington <bcodding@redhat.com>
Wed, 6 Feb 2019 11:09:43 +0000 (06:09 -0500)
committerAnna Schumaker <Anna.Schumaker@Netapp.com>
Tue, 12 Feb 2019 20:56:28 +0000 (15:56 -0500)
If nfs_page_async_flush() removes the page from the mapping, then we can't
use page_file_mapping() on it as nfs_updatepate() is wont to do when
receiving an error.  Instead, push the mapping to the stack before the page
is possibly truncated.

Fixes: 8fc75bed96bb ("NFS: Fix up return value on fatal errors in nfs_page_async_flush()")
Signed-off-by: Benjamin Coddington <bcodding@redhat.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
fs/nfs/write.c

index f12cb31a41e5ba5bb1b0bbc7d710cecb11cf4e69..d09c9f878141a5c4213aef1faccb85b979bf4793 100644 (file)
@@ -238,9 +238,9 @@ out:
 }
 
 /* A writeback failed: mark the page as bad, and invalidate the page cache */
-static void nfs_set_pageerror(struct page *page)
+static void nfs_set_pageerror(struct address_space *mapping)
 {
-       nfs_zap_mapping(page_file_mapping(page)->host, page_file_mapping(page));
+       nfs_zap_mapping(mapping->host, mapping);
 }
 
 /*
@@ -994,7 +994,7 @@ static void nfs_write_completion(struct nfs_pgio_header *hdr)
                nfs_list_remove_request(req);
                if (test_bit(NFS_IOHDR_ERROR, &hdr->flags) &&
                    (hdr->good_bytes < bytes)) {
-                       nfs_set_pageerror(req->wb_page);
+                       nfs_set_pageerror(page_file_mapping(req->wb_page));
                        nfs_context_set_write_error(req->wb_context, hdr->error);
                        goto remove_req;
                }
@@ -1348,7 +1348,8 @@ int nfs_updatepage(struct file *file, struct page *page,
                unsigned int offset, unsigned int count)
 {
        struct nfs_open_context *ctx = nfs_file_open_context(file);
-       struct inode    *inode = page_file_mapping(page)->host;
+       struct address_space *mapping = page_file_mapping(page);
+       struct inode    *inode = mapping->host;
        int             status = 0;
 
        nfs_inc_stats(inode, NFSIOS_VFSUPDATEPAGE);
@@ -1366,7 +1367,7 @@ int nfs_updatepage(struct file *file, struct page *page,
 
        status = nfs_writepage_setup(ctx, page, offset, count);
        if (status < 0)
-               nfs_set_pageerror(page);
+               nfs_set_pageerror(mapping);
        else
                __set_page_dirty_nobuffers(page);
 out: