kern: pts: do not special case closed slave side

This would previously return 1 if the slave side of the pts was closed
to force an application to read() from it and observe the EOF, but it's
not clear why and this is inconsistent both with how we handle devices
with similar mechanics (like pipes) and also with other kernels, such as
OpenBSD/NetBSD and Linux.

PR:             239604
Reviewed by:	kib

(cherry picked from commit 30189156d325fbcc9d1997d791daedc9fa3bed20)

3 files changed, 68 insertions, 6 deletions

diff --git a/sys/kern/tty_pts.c b/sys/kern/tty_pts.c
index f79c9cc3e979..31122d422a28 100644
--- a/sys/kern/tty_pts.c
+++ b/sys/kern/tty_pts.c
@@ -271,12 +271,7 @@ ptsdev_ioctl(struct file *fp, u_long cmd, void *data,
 		return (0);
 	case FIONREAD:
 		tty_lock(tp);
-		if (psc->pts_flags & PTS_FINISHED) {
-			/* Force read() to be called. */
-			*(int *)data = 1;
-		} else {
-			*(int *)data = ttydisc_getc_poll(tp);
-		}
+		*(int *)data = ttydisc_getc_poll(tp);
 		tty_unlock(tp);
 		return (0);
 	case FIODGNAME:
diff --git a/tests/sys/kern/Makefile b/tests/sys/kern/Makefile
index e449d9deeed1..b95be12d9665 100644
--- a/tests/sys/kern/Makefile
+++ b/tests/sys/kern/Makefile
@@ -38,6 +38,7 @@ ATF_TESTS_C+=	subr_physmem_test
 PLAIN_TESTS_C+=	subr_unit_test
 ATF_TESTS_C+=	sysctl_kern_proc
 ATF_TESTS_C+=	sys_getrandom
+ATF_TESTS_C+=	tty_pts
 ATF_TESTS_C+=	unix_dgram
 ATF_TESTS_C+=	unix_passfd_dgram
 TEST_METADATA.unix_passfd_dgram+=	is_exclusive="true"
@@ -86,6 +87,8 @@ NETBSD_ATF_TESTS_C+=	sysv_test
 CFLAGS.mqueue_test+=	-I${SRCTOP}/tests
 LIBADD.mqueue_test+=	rt
 
+LIBADD.tty_pts+=	atf_c util
+
 ATF_TESTS_C+=	libkern_crc32
 SRCS.libkern_crc32+=	libkern_crc32.c
 .PATH: ${SRCTOP}/sys/libkern
diff --git a/tests/sys/kern/tty_pts.c b/tests/sys/kern/tty_pts.c
new file mode 100644
index 000000000000..241cb085a5a7
--- /dev/null
+++ b/tests/sys/kern/tty_pts.c
@@ -0,0 +1,64 @@
+/*-
+ * Copyright (c) 2024 Kyle Evans <kevans@FreeBSD.org>
+ *
+ * SPDX-License-Identifier: BSD-2-Clause
+ */
+
+#include <sys/types.h>
+#include <sys/ioctl.h>
+
+#include <termios.h>
+
+#include <atf-c.h>
+#include <libutil.h>
+
+/* Just a little more concise. */
+#define	newpty(masterp, slavep)	openpty((masterp), (slavep), NULL, NULL, NULL)
+
+ATF_TC_WITHOUT_HEAD(fionread);
+ATF_TC_BODY(fionread, tc)
+{
+	char rbuf[32];
+	char buf[] = "Hello";
+	int master, slave;
+	int bytes;
+
+	ATF_REQUIRE_EQ(0, newpty(&master, &slave));
+
+	/* Should be empty to begin with. */
+	ATF_REQUIRE_EQ(0, ioctl(master, FIONREAD, &bytes));
+	ATF_REQUIRE_EQ(0, bytes);
+
+	ATF_REQUIRE_EQ(sizeof(buf) - 1, write(slave, buf, sizeof(buf) - 1));
+	ATF_REQUIRE_EQ(0, ioctl(master, FIONREAD, &bytes));
+	ATF_REQUIRE_EQ(sizeof(buf) - 1, bytes);
+
+	/* Drain what we have available, should result in 0 bytes again. */
+	ATF_REQUIRE_EQ(sizeof(buf) - 1, read(master, rbuf, sizeof(rbuf)));
+	ATF_REQUIRE_EQ(0, ioctl(master, FIONREAD, &bytes));
+	ATF_REQUIRE_EQ(0, bytes);
+
+	/*
+	 * Write once more, then close the slave side with data still in the
+	 * buffer.
+	 */
+	ATF_REQUIRE_EQ(sizeof(buf) - 1, write(slave, buf, sizeof(buf) - 1));
+	ATF_REQUIRE_EQ(0, ioctl(master, FIONREAD, &bytes));
+	ATF_REQUIRE_EQ(sizeof(buf) - 1, bytes);
+
+	ATF_REQUIRE_EQ(0, close(slave));
+
+	/*
+	 * The tty's output queue is discarded upon close, so we shouldn't have
+	 * anything else to read().
+	 */
+	ATF_REQUIRE_EQ(0, ioctl(master, FIONREAD, &bytes));
+	ATF_REQUIRE_EQ(0, bytes);
+	ATF_REQUIRE_EQ(0, read(master, rbuf, sizeof(rbuf)));
+}
+
+ATF_TP_ADD_TCS(tp)
+{
+	ATF_TP_ADD_TC(tp, fionread);
+	return (atf_no_error());
+}