Skip to contentSkip to conversation

igt/vc4_dmabuf_poll: Add a test for polling to wait for dmabuf fences.

PATCH In review Eric Anholt 2 participants 3 messages Apr 11, 2017 – Jun 7, 2017

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 igt/vc4_dmabuf_poll: Add a test for polling to wait for dmabuf fences.

This successfully catches vc4's lack of dmabuf fencing.

---

Has anyone looked into shared infrastructure for tests to do KMS/dmabuf/etc. things with a generic "get a BO that's being rendered to for this driver" call?

 tests/Makefile.am       |  2 ++
 tests/Makefile.sources  |  1 +
 tests/vc4_dmabuf_poll.c | 85 +++++++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 88 insertions(+)
 create mode 100644 tests/vc4_dmabuf_poll.c
tests/Makefile.am +2 -0
index 8930c245043d..7c40c9ea8276 100644
@@ -114,6 +114,8 @@ vc4_create_bo_CFLAGS = $(AM_CFLAGS) $(DRM_VC4_CFLAGS)
 vc4_create_bo_LDADD = $(LDADD) $(DRM_VC4_LIBS)
 vc4_lookup_fail_CFLAGS = $(AM_CFLAGS) $(DRM_VC4_CFLAGS)
 vc4_lookup_fail_LDADD = $(LDADD) $(DRM_VC4_LIBS)
+vc4_dmabuf_poll_CFLAGS = $(AM_CFLAGS) $(DRM_VC4_CFLAGS)
+vc4_dmabuf_poll_LDADD = $(LDADD) $(DRM_VC4_LIBS)
 vc4_wait_bo_CFLAGS = $(AM_CFLAGS) $(DRM_VC4_CFLAGS)
 vc4_wait_bo_LDADD = $(LDADD) $(DRM_VC4_LIBS)
 vc4_wait_seqno_CFLAGS = $(AM_CFLAGS) $(DRM_VC4_CFLAGS)
tests/Makefile.sources +1 -0
index 6e07d9387f83..b579e28c6669 100644
@@ -6,6 +6,7 @@ NOUVEAU_TESTS_M = \
 
 VC4_TESTS_M = \
 	vc4_create_bo \
+	vc4_dmabuf_poll \
 	vc4_lookup_fail \
 	vc4_wait_bo \
 	vc4_wait_seqno \
tests/vc4_dmabuf_poll.c new file +85 -0

85 lines not on this page · raw message

Signature · 6 lines
2.11.0

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#1 lore raw
Quoted 8 lines · Eric Anholt

This successfully catches vc4's lack of dmabuf fencing.

---

Has anyone looked into shared infrastructure for tests to do KMS/dmabuf/etc. things with a generic "get a BO that's being rendered to for this driver" call?

We have some helpers for i915 to make a bo busy with an explicit release (we create a looping batch which we break with a cpu write), for extremely well-control busy tests. Not sure how well that's portable, and without full control it's hard to make busy tests reliable.
-Daniel

Quoted 130 lines
 tests/Makefile.am       |  2 ++
 tests/Makefile.sources  |  1 +
 tests/vc4_dmabuf_poll.c | 85 +++++++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 88 insertions(+)
 create mode 100644 tests/vc4_dmabuf_poll.c
tests/Makefile.am +2 -0
index 8930c245043d..7c40c9ea8276 100644
@@ -114,6 +114,8 @@ vc4_create_bo_CFLAGS = $(AM_CFLAGS) $(DRM_VC4_CFLAGS)
 vc4_create_bo_LDADD = $(LDADD) $(DRM_VC4_LIBS)
 vc4_lookup_fail_CFLAGS = $(AM_CFLAGS) $(DRM_VC4_CFLAGS)
 vc4_lookup_fail_LDADD = $(LDADD) $(DRM_VC4_LIBS)
+vc4_dmabuf_poll_CFLAGS = $(AM_CFLAGS) $(DRM_VC4_CFLAGS)
+vc4_dmabuf_poll_LDADD = $(LDADD) $(DRM_VC4_LIBS)
 vc4_wait_bo_CFLAGS = $(AM_CFLAGS) $(DRM_VC4_CFLAGS)
 vc4_wait_bo_LDADD = $(LDADD) $(DRM_VC4_LIBS)
 vc4_wait_seqno_CFLAGS = $(AM_CFLAGS) $(DRM_VC4_CFLAGS)
tests/Makefile.sources +1 -0
index 6e07d9387f83..b579e28c6669 100644
@@ -6,6 +6,7 @@ NOUVEAU_TESTS_M = \
 
 VC4_TESTS_M = \
 	vc4_create_bo \
+	vc4_dmabuf_poll \
 	vc4_lookup_fail \
 	vc4_wait_bo \
 	vc4_wait_seqno \
tests/vc4_dmabuf_poll.c new file +85 -0
new file mode 100644
index 000000000000..ac44d4cf80c7
@@ -0,0 +1,85 @@
+/*
+ * Copyright © 2017 Broadcom
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include "igt.h"
+#include "igt_vc4.h"
+#include <unistd.h>
+#include <stdlib.h>
+#include <stdio.h>
+#include <string.h>
+#include <fcntl.h>
+#include <inttypes.h>
+#include <errno.h>
+#include <sys/stat.h>
+#include <sys/ioctl.h>
+#include <poll.h>
+#include "vc4_drm.h"
+
+static void
+poll_write_bo_test(int fd, int poll_flag)
+{
+	size_t size = 1024 * 1024 * 4;
+	uint32_t clearval = 0xaabbccdd;
+	/* Get a BO that's being rendered to. */
+	int handle = igt_vc4_get_cleared_bo(fd, size, clearval);
+	int dmabuf_fd = prime_handle_to_fd(fd, handle);
+	struct pollfd p = {
+		.fd = dmabuf_fd,
+		.events = POLLIN,
+	};
+	struct drm_vc4_wait_bo wait = {
+		.handle = handle,
+		.timeout_ns = 0,
+	};
+
+	/* Block for a couple of minutes waiting for rendering to complete. */
+	int poll_ret = poll(&p, 1, 120 * 1000);
+	igt_assert(poll_ret == 1);
+
+	/* Now that we've waited for idle, a nonblocking wait for the
+	 * BO should pass.
+	 */
+	do_ioctl(fd, DRM_IOCTL_VC4_WAIT_BO, &wait);
+
+	close(dmabuf_fd);
+	gem_close(fd, handle);
+}
+
+igt_main
+{
+	int fd;
+
+	igt_fixture
+		fd = drm_open_driver(DRIVER_VC4);
+
+	igt_subtest("poll-write-waits-until-write-done") {
+		poll_write_bo_test(fd, POLLOUT);
+	}
+
+	igt_subtest("poll-read-waits-until-write-done") {
+		poll_write_bo_test(fd, POLLIN);
+	}
+
+	igt_fixture
+		close(fd);
+}
Signature · 6 lines
2.11.0

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#2 lore raw
Quoted 6 lines · Daniel Vetter
Quoted 5 lines · Eric Anholt

This successfully catches vc4's lack of dmabuf fencing.

---

Quoted 8 lines
Quoted 3 lines

Has anyone looked into shared infrastructure for tests to do KMS/dmabuf/etc. things with a generic "get a BO that's being rendered to for this driver" call?

We have some helpers for i915 to make a bo busy with an explicit release (we create a looping batch which we break with a cpu write), for extremely well-control busy tests. Not sure how well that's portable, and without full control it's hard to make busy tests reliable.

I think I would need a new kernel ioctl for that.

I can queue up 32MB read/write jobs, and with ~1GB/s of memory bandwidth, that's been plenty of time to write tests that catch at least some bugs in synchronization.

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx