Received-SPF: pass (google.com: domain of linux-nfs-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Subject: [PATCH] nfsv4.0/read: Test the behavior of reading near OFFSET_MAX
From:   Chuck Lever <chuck.lever@oracle.com>
To:     bfields@fieldses.org
Cc:     linux-nfs@vger.kernel.org
Date:   Sat, 05 Feb 2022 12:05:29 -0500
Message-ID: <164408072927.1028772.2263116854233914910.stgit@morisot.1015granger.net>
User-Agent: StGit/1.5
MIME-Version: 1.0
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: 7bit
Precedence: bulk

On many systems, the internal representation of a file offset or
file size is a signed 64-bit value. NFS, however, uses an unsigned
value for these quantities. The server must convert incoming offsets
and file sizes properly or risk an underflow.

Add a test which exercises this corner case.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
---
 nfs4.0/servertests/st_read.py |   12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/nfs4.0/servertests/st_read.py b/nfs4.0/servertests/st_read.py
index 1b27f5d06f2f..f75b753d61c7 100644
--- a/nfs4.0/servertests/st_read.py
+++ b/nfs4.0/servertests/st_read.py
@@ -91,6 +91,18 @@ def testLargeOffset(t, env):
     check(res, msg="Reading file /%s" % b'/'.join(env.opts.usefile))
     _compare(t, res, b'', True)
 
+def testVeryLargeOffset(t, env):
+    """READ with offset far outside file
+
+    FLAGS: read all
+    DEPEND: LOOKFILE
+    CODE: RD5a
+    """
+    c = env.c1
+    res = c.read_file(env.opts.usefile, 0x7ffffffffffffffc, 10)
+    check(res, msg="Reading file /%s" % b'/'.join(env.opts.usefile))
+    _compare(t, res, b'', True)
+
 def testZeroCount(t, env):
     """READ with count=0