Linux NFS development
 help / color / mirror / Atom feed
From: Jeff Layton <jlayton@kernel.org>
To: Calum Mackay <calum.mackay@oracle.com>
Cc: Chuck Lever <cel@kernel.org>, NeilBrown <neil@brown.name>,
	 Olga Kornievskaia <okorniev@redhat.com>,
	Dai Ngo <Dai.Ngo@oracle.com>,  Tom Talpey <tom@talpey.com>,
	linux-nfs@vger.kernel.org,  Jeff Layton <jlayton@kernel.org>
Subject: [PATCH pynfs 01/13] server41tests: add helpers and basic synchronous COPY test
Date: Thu, 09 Jul 2026 15:02:29 -0400	[thread overview]
Message-ID: <20260709-copy-v1-1-849bf581d7cb@kernel.org> (raw)
In-Reply-To: <20260709-copy-v1-0-849bf581d7cb@kernel.org>

Add helpers to reduce boilerplate in COPY tests: _do_copy(),
_create_and_open(), _poll_offload_status(), and chunked _write_data()/
_verify_data() that split I/O into pieces bounded by the session's
negotiated ca_maxrequestsize/ca_maxresponsesize (a single large WRITE or
READ would otherwise fail with NFS4ERR_REQ_TOO_BIG).

Add testSyncCopy (COPY1): write 64KB to a source file, copy it, and
verify the destination contents.  RFC 7862 permits a server to perform
the copy asynchronously even when a synchronous copy is requested, so
honor cr_resok4.cr_requirements.cr_synchronous: poll OFFLOAD_STATUS to
completion when the server downgrades to async instead of failing on a
zero wr_count.

Signed-off-by: Jeff Layton <jlayton@kernel.org>
---
 nfs4.1/server41tests/st_copy.py | 90 ++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 89 insertions(+), 1 deletion(-)

diff --git a/nfs4.1/server41tests/st_copy.py b/nfs4.1/server41tests/st_copy.py
index a4bdb77ca407..c7e48adf6fbe 100644
--- a/nfs4.1/server41tests/st_copy.py
+++ b/nfs4.1/server41tests/st_copy.py
@@ -1,14 +1,102 @@
+import time
+
 from .st_create_session import create_session
 from xdrdef.nfs4_const import *
 
 from .environment import check, fail, create_file, open_file, close_file
-from .environment import open_create_file_op, use_obj, write_file
+from .environment import open_create_file_op, use_obj, write_file, read_file
 from xdrdef.nfs4_type import open_owner4, openflag4, createhow4, open_claim4
 from xdrdef.nfs4_type import creatverfattr, fattr4, stateid4, locker4, lock_owner4
 from xdrdef.nfs4_type import open_to_lock_owner4
 import nfs_ops
 op = nfs_ops.NFS4ops()
 
+def _do_copy(sess, src_fh, src_stateid, dst_fh, dst_stateid,
+             src_offset=0, dst_offset=0, count=0,
+             consecutive=0, synchronous=1):
+    ops = [op.putfh(src_fh), op.savefh(), op.putfh(dst_fh),
+           op.copy(src_stateid, dst_stateid, src_offset, dst_offset,
+                   count, consecutive, synchronous, [])]
+    return sess.compound(ops)
+
+def _poll_offload_status(sess, dst_fh, copy_stateid, timeout=120):
+    deadline = time.time() + timeout
+    while time.time() < deadline:
+        ops = [op.putfh(dst_fh), op.offload_status(copy_stateid)]
+        res = sess.compound(ops)
+        check(res)
+        status_res = res.resarray[-1]
+        if status_res.osr_complete:
+            return status_res
+        time.sleep(1)
+    fail("OFFLOAD_STATUS did not complete within %d seconds" % timeout)
+
+def _create_and_open(sess, name):
+    res = create_file(sess, name)
+    check(res)
+    fh = res.resarray[-1].object
+    stateid = res.resarray[-2].stateid
+    return fh, stateid
+
+def _write_data(sess, fh, stateid, data, offset=0):
+    """Write data in chunks bounded by the session's max request size."""
+    chunk = sess.fore_channel.maxrequestsize - 1024
+    pos = 0
+    while pos < len(data):
+        res = write_file(sess, fh, data[pos:pos + chunk], offset + pos, stateid)
+        check(res, msg="WRITE at offset %d" % (offset + pos))
+        pos += res.count
+
+def _verify_data(sess, fh, stateid, data, offset=0):
+    """Read back and compare data in chunks bounded by max response size."""
+    chunk = sess.fore_channel.maxresponsesize - 1024
+    pos = 0
+    while pos < len(data):
+        res = read_file(sess, fh, offset + pos, min(chunk, len(data) - pos),
+                        stateid)
+        check(res)
+        if not res.data:
+            fail("Short read at offset %d" % (offset + pos))
+        if res.data != data[pos:pos + len(res.data)]:
+            fail("Data mismatch at offset %d" % (offset + pos))
+        pos += len(res.data)
+
+def testSyncCopy(t, env):
+    """synchronous copy of a file and verify contents
+
+    FLAGS: copy
+    CODE: COPY1
+    """
+    sess = env.c1.new_client_session(env.testname(t))
+    src_fh, src_stateid = _create_and_open(sess, env.testname(t))
+    data = b"A" * 65536
+    _write_data(sess, src_fh, src_stateid, data)
+
+    dst_fh, dst_stateid = _create_and_open(sess, env.testname(t) + b"_dst")
+
+    res = _do_copy(sess, src_fh, src_stateid, dst_fh, dst_stateid,
+                   count=len(data), synchronous=1)
+    check(res)
+    cr = res.resarray[-1]
+
+    # A synchronous COPY was requested, but RFC 7862 permits the server to
+    # perform the copy asynchronously anyway; cr_requirements.cr_synchronous
+    # reports what actually happened.  Honor either, but verify the byte count.
+    if cr.cr_resok4.cr_requirements.cr_synchronous:
+        if cr.cr_response.wr_count != len(data):
+            fail("Synchronous copy expected %d bytes, got %d" %
+                 (len(data), cr.cr_response.wr_count))
+    else:
+        copy_stateid = cr.cr_response.wr_callback_id[0]
+        status = _poll_offload_status(sess, dst_fh, copy_stateid)
+        if status.osr_complete[0] != NFS4_OK:
+            fail("Async copy completed with error: %d" % status.osr_complete[0])
+        if status.osr_count != len(data):
+            fail("Expected %d bytes copied, got %d" %
+                 (len(data), status.osr_count))
+
+    _verify_data(sess, dst_fh, dst_stateid, data)
+
 def testZeroLengthCopy(t, env):
     """test that zero-length copy copies to EOF
 

-- 
2.55.0


  reply	other threads:[~2026-07-09 19:02 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-07-09 19:02 [PATCH pynfs 00/13] server41tests: add some tests for copy offload Jeff Layton
2026-07-09 19:02 ` Jeff Layton [this message]
2026-07-09 19:02 ` [PATCH pynfs 02/13] server41tests: test COPY with non-zero offsets Jeff Layton
2026-07-09 19:02 ` [PATCH pynfs 03/13] server41tests: test async COPY with OFFLOAD_STATUS polling Jeff Layton
2026-07-09 19:02 ` [PATCH pynfs 04/13] server41tests: test OFFLOAD_STATUS persists after async copy completes Jeff Layton
2026-07-09 19:02 ` [PATCH pynfs 05/13] server41tests: test OFFLOAD_CANCEL on async copy Jeff Layton
2026-07-09 19:02 ` [PATCH pynfs 06/13] server41tests: test COPY with bad source stateid Jeff Layton
2026-07-09 19:02 ` [PATCH pynfs 07/13] server41tests: test COPY with bad destination stateid Jeff Layton
2026-07-09 19:02 ` [PATCH pynfs 08/13] server41tests: test OFFLOAD_STATUS with fabricated stateid Jeff Layton
2026-07-09 19:02 ` [PATCH pynfs 09/13] server41tests: test COPY within same file Jeff Layton
2026-07-09 19:02 ` [PATCH pynfs 10/13] nfs4.1: fix COPY_NOTIFY args union arm name in XDR Jeff Layton
2026-07-09 19:02 ` [PATCH pynfs 11/13] server41tests: test COPY_NOTIFY Jeff Layton
2026-07-09 19:02 ` [PATCH pynfs 12/13] server41tests: support a second server for inter-server copy Jeff Layton
2026-07-09 19:02 ` [PATCH pynfs 13/13] server41tests: test inter-server COPY Jeff Layton

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20260709-copy-v1-1-849bf581d7cb@kernel.org \
    --to=jlayton@kernel.org \
    --cc=Dai.Ngo@oracle.com \
    --cc=calum.mackay@oracle.com \
    --cc=cel@kernel.org \
    --cc=linux-nfs@vger.kernel.org \
    --cc=neil@brown.name \
    --cc=okorniev@redhat.com \
    --cc=tom@talpey.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox