Writes held before the shell's clock is first seen no longer overflow its first step

A 9P write served before a detached host's first pump holds its batch stamped at time 0, so nextWake said to wake at 20 ms while advance, seeing the clock for the first time, set stepped_ns days later and counted due - stepped_ns back past zero: an integer-overflow panic. It was what the 64 KiB msize trial hit in fs-test (restore detached space, the connection closing at the first write's clunk), a race any msize can lose. A batch's wake is now never before the next frame.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

1 files changed, 15 insertions, 0 deletions

diff --git a/src/ninep/pane.zig b/src/ninep/pane.zig
index 4327c403..bb4ebf86 100644
--- a/src/ninep/pane.zig
+++ b/src/ninep/pane.zig
@@ -1190,6 +1190,21 @@ test "an open's writes in a row are held and go in as one edit, seen by the next
     try testing.expectEqualStrings("abone\n", rd(p, body, 0, 64).bytes);
 }
 
+test "writes held before the shell's clock is first seen go in at its first step" {
+    const p = try withFile(testing.allocator, "one\n");
+    defer p.deinit();
+    const body = Node.of(serialOf(p), .body);
+    try testing.expect(!p.clock_started);
+    const h = call(p, .{ .tag = 1, .op = .open, .node = body, .omode = 1 }).reply.handle;
+    const revision = p.panes[0].?.file.?.revision;
+    _ = call(p, .{ .tag = 2, .op = .write, .node = body, .handle = h, .data = "x\n" });
+    try testing.expectEqual(@as(u64, 0), p.fs.batch.last_ns);
+    // A detached host's first step: its clock is days into the machine's.
+    p.advance(700_000 * std.time.ns_per_s);
+    flushPausedBatch(p);
+    try testing.expectEqual(revision + 1, p.panes[0].?.file.?.revision);
+}
+
 test "a tag write past the limit is refused whole, naming the limit" {
     const p = try withFile(testing.allocator, "x\n");
     defer p.deinit();