[RFC,i-g-t,v2] tests/gem_exec_pad_to_size: Test object padding at execbuf

Submitted by Tvrtko Ursulin on April 1, 2015, 11:21 a.m.

Details

Message ID 1427887274-29195-1-git-send-email-tvrtko.ursulin@linux.intel.com
State New
Headers show

Not browsing as part of any series.

Commit Message

Tvrtko Ursulin April 1, 2015, 11:21 a.m.
From: Tvrtko Ursulin <tvrtko.ursulin@intel.com>

This tests the new EXEC_OBJECT_PAD_TO_SIZE exec_object2 flag.

Similar to some other tests, it uses knowledge of the DRM
allocation policy in order to get two objects mapped adjacent
to each other. It is then possible to verify that the pad to
size flag will move them apart.

v2: Correct commit message. (Chris Wilson)

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
---
 tests/Makefile.sources       |   1 +
 tests/gem_exec_pad_to_size.c | 276 +++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 277 insertions(+)
 create mode 100644 tests/gem_exec_pad_to_size.c

Patch hide | download patch | download mbox

diff --git a/tests/Makefile.sources b/tests/Makefile.sources
index 0a974a6..5f21728 100644
--- a/tests/Makefile.sources
+++ b/tests/Makefile.sources
@@ -34,6 +34,7 @@  TESTS_progs_M = \
 	gem_exec_bad_domains \
 	gem_exec_faulting_reloc \
 	gem_exec_nop \
+	gem_exec_pad_to_size \
 	gem_exec_params \
 	gem_exec_parse \
 	gem_fenced_exec_thrash \
diff --git a/tests/gem_exec_pad_to_size.c b/tests/gem_exec_pad_to_size.c
new file mode 100644
index 0000000..3584e5c
--- /dev/null
+++ b/tests/gem_exec_pad_to_size.c
@@ -0,0 +1,276 @@ 
+/*
+ * Copyright © 2015 Intel Corporation
+ *
+ * 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.
+ *
+ * Authors:
+ *    Tvrtko Ursulin <tvrtko.ursulin@intel.com>
+ *
+ */
+
+#include <unistd.h>
+#include <stdlib.h>
+#include <stdint.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 <sys/time.h>
+#include "drm.h"
+#include "ioctl_wrappers.h"
+#include "igt_core.h"
+#include "drmtest.h"
+#include "intel_reg.h"
+
+#ifndef PAGE_SIZE
+#define PAGE_SIZE 4096
+#endif
+
+struct local_drm_i915_gem_exec_object2 {
+	/**
+	 * User's handle for a buffer to be bound into the GTT for this
+	 * operation.
+	 */
+	__u32 handle;
+
+	/** Number of relocations to be performed on this buffer */
+	__u32 relocation_count;
+	/**
+	 * Pointer to array of struct drm_i915_gem_relocation_entry containing
+	 * the relocations to be performed in this buffer.
+	 */
+	__u64 relocs_ptr;
+
+	/** Required alignment in graphics aperture */
+	__u64 alignment;
+
+	/**
+	 * Returned value of the updated offset of the object, for future
+	 * presumed_offset writes.
+	 */
+	__u64 offset;
+
+#define LOCAL_EXEC_OBJECT_NEEDS_FENCE (1<<0)
+#define LOCAL_EXEC_OBJECT_NEEDS_GTT	(1<<1)
+#define LOCAL_EXEC_OBJECT_WRITE	(1<<2)
+#define LOCAL_EXEC_OBJECT_PAD_TO_SIZE (1<<3)
+#define __LOCAL_EXEC_OBJECT_UNKNOWN_FLAGS -(LOCAL_EXEC_OBJECT_PAD_TO_SIZE<<1)
+	__u64 flags;
+
+	__u64 pad_to_size;
+	__u64 rsvd2;
+};
+
+static int
+exec(int fd, uint32_t handles[3], uint32_t pad_to_size[2], uint64_t *offsets)
+{
+	struct drm_i915_gem_execbuffer2 execbuf;
+	struct local_drm_i915_gem_exec_object2 gem_exec[3];
+	int ret = 0;
+
+	memset(gem_exec, 0, sizeof(gem_exec));
+
+	gem_exec[0].handle = handles[1];
+	gem_exec[0].relocation_count = 0;
+	gem_exec[0].relocs_ptr = 0;
+	gem_exec[0].alignment = 0;
+	gem_exec[0].offset = 0;
+	gem_exec[0].flags = pad_to_size[0] ?
+			    LOCAL_EXEC_OBJECT_PAD_TO_SIZE : 0;
+	gem_exec[0].pad_to_size = pad_to_size[0];
+	gem_exec[0].rsvd2 = 0;
+
+	gem_exec[1].handle = handles[2];
+	gem_exec[1].relocation_count = 0;
+	gem_exec[1].relocs_ptr = 0;
+	gem_exec[1].alignment = 0;
+	gem_exec[1].offset = 0;
+	gem_exec[1].flags = pad_to_size[1] ?
+			    LOCAL_EXEC_OBJECT_PAD_TO_SIZE : 0;
+	gem_exec[1].pad_to_size = pad_to_size[1];
+	gem_exec[1].rsvd2 = 0;
+
+	gem_exec[2].handle = handles[0];
+	gem_exec[2].relocation_count = 0;
+	gem_exec[2].relocs_ptr = 0;
+	gem_exec[2].alignment = 0;
+	gem_exec[2].offset = 0;
+	gem_exec[2].flags = 0;
+	gem_exec[2].pad_to_size = 0;
+	gem_exec[2].rsvd2 = 0;
+
+	execbuf.buffers_ptr = (uintptr_t)gem_exec;
+	execbuf.buffer_count = 3;
+	execbuf.batch_start_offset = 0;
+	execbuf.batch_len = 8;
+	execbuf.cliprects_ptr = 0;
+	execbuf.num_cliprects = 0;
+	execbuf.DR1 = 0;
+	execbuf.DR4 = 0;
+	execbuf.flags = I915_EXEC_RENDER;
+	i915_execbuffer2_set_context_id(execbuf, 0);
+	execbuf.rsvd2 = 0;
+
+	if (drmIoctl(fd, DRM_IOCTL_I915_GEM_EXECBUFFER2, &execbuf))
+			ret = -errno;
+
+	if (ret == 0) {
+		gem_sync(fd, handles[0]);
+
+		if (offsets) {
+			offsets[0] = gem_exec[0].offset;
+			offsets[1] = gem_exec[1].offset;
+		}
+	}
+
+	return ret;
+}
+
+static void
+require_pad_to_size(int fd, uint32_t handles[3])
+{
+	igt_require(exec(fd, handles, (uint32_t[2]){ PAGE_SIZE, PAGE_SIZE },
+			 NULL) == 0);
+}
+
+static void
+not_aligned(int fd, uint32_t handles[3])
+{
+	require_pad_to_size(fd, handles);
+
+	igt_require(exec(fd, handles, (uint32_t[2]){1 ,1},
+			 NULL) == -EINVAL);
+}
+
+static void
+padding(int fd, uint32_t handles[3])
+{
+	uint32_t pad_to_size[2] = {0, 0};
+	uint64_t offsets[2];
+	uint32_t distance;
+	bool neighbours = false;
+	unsigned int try, idx;
+	const unsigned int max_tries = 1024; /* Finger in the air. */
+	uint32_t *loc_handles;
+	uint32_t eb_handles[3];
+
+	require_pad_to_size(fd, handles);
+
+	loc_handles = calloc(max_tries * 2, sizeof(uint32_t));
+	igt_assert(loc_handles);
+
+	/* Try with passed in handles first. */
+	loc_handles[0] = handles[1];
+	loc_handles[1] = handles[2];
+
+	/* Try to get two buffer object next to each other in GTT space. */
+	for (try = 0, idx = 0; try < max_tries;) {
+		eb_handles[0] = handles[0];
+		eb_handles[1] = loc_handles[idx];
+		eb_handles[2] = loc_handles[idx + 1];
+
+		igt_assert(exec(fd, eb_handles, (uint32_t[2]){0, 0},
+				offsets) == 0);
+
+		if (offsets[1] > offsets[0]) {
+			distance = offsets[1] - offsets[0];
+			if (distance == PAGE_SIZE)
+				neighbours = true;
+			pad_to_size[0] = ALIGN(distance + PAGE_SIZE, PAGE_SIZE);
+		} else {
+			distance = offsets[0] - offsets[1];
+			if (distance == PAGE_SIZE)
+				neighbours = true;
+			pad_to_size[1] = ALIGN(distance + PAGE_SIZE, PAGE_SIZE);
+		}
+
+		if (neighbours)
+			break;
+
+		try++;
+		idx +=2;
+
+		loc_handles[idx] = gem_create(fd, PAGE_SIZE);
+		igt_assert(loc_handles[idx]);
+		loc_handles[idx + 1] = gem_create(fd, PAGE_SIZE);
+		igt_assert(loc_handles[idx + 1]);
+	}
+
+	/* Test can't confidently run. */
+	if (!neighbours)
+		goto cleanup;
+
+	/* Re-exec with padding set. */
+	igt_assert(exec(fd, eb_handles, pad_to_size, offsets) == 0);
+
+	if (offsets[1] > offsets[0])
+		distance = offsets[1] - offsets[0];
+	else
+		distance = offsets[0] - offsets[1];
+
+	/* Check that object are now away from each other. */
+	igt_assert(distance >= 2 * PAGE_SIZE);
+
+cleanup:
+	/* Cleanup our bos. */
+	for (idx = 0; loc_handles[idx] && (idx < try * 2); idx++)
+		gem_close(fd, loc_handles[2 + idx]);
+
+	free(loc_handles);
+
+	igt_require(neighbours);
+}
+
+uint32_t batch[2] = {MI_BATCH_BUFFER_END};
+uint32_t handles[3];
+int fd;
+
+igt_main
+{
+	igt_fixture {
+		fd = drm_open_any();
+
+		handles[0] = gem_create(fd, PAGE_SIZE);
+		gem_write(fd, handles[0], 0, batch, sizeof(batch));
+
+		handles[1] = gem_create(fd, PAGE_SIZE);
+		handles[2] = gem_create(fd, PAGE_SIZE);
+	}
+
+	igt_subtest("pad_to_size")
+		require_pad_to_size(fd, handles);
+
+	igt_subtest("not_aligned")
+		not_aligned(fd, handles);
+
+	igt_subtest("padding")
+		padding(fd, handles);
+
+	igt_fixture {
+		gem_close(fd, handles[0]);
+		gem_close(fd, handles[1]);
+		gem_close(fd, handles[2]);
+
+		close(fd);
+	}
+}

Comments

On Wed, Apr 01, 2015 at 12:21:14PM +0100, Tvrtko Ursulin wrote:

> +static int
> +exec(int fd, uint32_t handles[3], uint32_t pad_to_size[2], uint64_t *offsets)
> +{
> +	struct drm_i915_gem_execbuffer2 execbuf;
> +	struct local_drm_i915_gem_exec_object2 gem_exec[3];
> +	int ret = 0;
> +
> +	memset(gem_exec, 0, sizeof(gem_exec));
> +
> +	gem_exec[0].handle = handles[1];
> +	gem_exec[0].relocation_count = 0;
> +	gem_exec[0].relocs_ptr = 0;
> +	gem_exec[0].alignment = 0;
> +	gem_exec[0].offset = 0;
Ignore the unused fields (they are explicitly memset(0)) so that we can
focus on the important ones under test.

> +	gem_exec[0].flags = pad_to_size[0] ?
> +			    LOCAL_EXEC_OBJECT_PAD_TO_SIZE : 0;
> +	gem_exec[0].pad_to_size = pad_to_size[0];
> +

memset(execbuf) and skip boring fields
> +	execbuf.buffers_ptr = (uintptr_t)gem_exec;
> +	execbuf.buffer_count = 3;
> +	execbuf.batch_start_offset = 0;
> +	execbuf.batch_len = 8;
> +	execbuf.flags = I915_EXEC_RENDER;
This can be just I915_EXEC_DEFAULT (i.e. 0);


> +	if (drmIoctl(fd, DRM_IOCTL_I915_GEM_EXECBUFFER2, &execbuf))
> +			ret = -errno;

> +	if (ret == 0) {
> +		gem_sync(fd, handles[0]);
Not required for this test. However... You probably want to do the
gem_sync() first. (Yes, there is an amusing reason to do :)

> +		if (offsets) {
> +			offsets[0] = gem_exec[0].offset;
> +			offsets[1] = gem_exec[1].offset;
> +		}
> +	}
> +
> +	return ret;
> +}
> +
> +static void
> +require_pad_to_size(int fd, uint32_t handles[3])
> +{
> +	igt_require(exec(fd, handles, (uint32_t[2]){ PAGE_SIZE, PAGE_SIZE },
> +			 NULL) == 0);
> +}
> +
> +static void
> +not_aligned(int fd, uint32_t handles[3])
> +{
> +	require_pad_to_size(fd, handles);
> +
> +	igt_require(exec(fd, handles, (uint32_t[2]){1 ,1},
> +			 NULL) == -EINVAL);
> +}
> +
> +static void
> +padding(int fd, uint32_t handles[3])
> +{
> +	uint32_t pad_to_size[2] = {0, 0};
> +	uint64_t offsets[2];
> +	uint32_t distance;
> +	bool neighbours = false;
> +	unsigned int try, idx;
> +	const unsigned int max_tries = 1024; /* Finger in the air. */
> +	uint32_t *loc_handles;
> +	uint32_t eb_handles[3];
> +
> +	require_pad_to_size(fd, handles);
> +
> +	loc_handles = calloc(max_tries * 2, sizeof(uint32_t));
> +	igt_assert(loc_handles);
> +
> +	/* Try with passed in handles first. */
> +	loc_handles[0] = handles[1];
> +	loc_handles[1] = handles[2];
> +
> +	/* Try to get two buffer object next to each other in GTT space. */
/* Try to get two buffer object next to each other in GTT space.
 * We presume that sequential reservations are likely to be aligned and
 * try until we find a pair that is.
 */

> +	for (try = 0, idx = 0; try < max_tries;) {
> +		eb_handles[0] = handles[0];
> +		eb_handles[1] = loc_handles[idx];
> +		eb_handles[2] = loc_handles[idx + 1];
> +
> +		igt_assert(exec(fd, eb_handles, (uint32_t[2]){0, 0},
> +				offsets) == 0);
> +
> +		if (offsets[1] > offsets[0]) {
> +			distance = offsets[1] - offsets[0];
> +			if (distance == PAGE_SIZE)
> +				neighbours = true;
> +			pad_to_size[0] = ALIGN(distance + PAGE_SIZE, PAGE_SIZE);
> +		} else {
> +			distance = offsets[0] - offsets[1];
> +			if (distance == PAGE_SIZE)
> +				neighbours = true;
> +			pad_to_size[1] = ALIGN(distance + PAGE_SIZE, PAGE_SIZE);
> +		}
> +
> +		if (neighbours)
> +			break;
> +
> +		try++;
> +		idx +=2;

Just use idx++ here and allocate a new handle one at a time. Just as
likely to be adjacent to the previous handle as the next one will be to
us. For extra paranoia, you could even try an evict-everything pass :)

Otherwise looks fine.
-Chris
On Wed, Apr 01, 2015 at 02:06:53PM +0100, Chris Wilson wrote:
> Just use idx++ here and allocate a new handle one at a time. Just as
> likely to be adjacent to the previous handle as the next one will be to
> us. For extra paranoia, you could even try an evict-everything pass :)

Which is as easy as igt_drop_caches_set(DROP_ALL);
-Chris
On 04/01/2015 02:06 PM, Chris Wilson wrote:
> On Wed, Apr 01, 2015 at 12:21:14PM +0100, Tvrtko Ursulin wrote:
>> +	if (drmIoctl(fd, DRM_IOCTL_I915_GEM_EXECBUFFER2, &execbuf))
>> +			ret = -errno;
>
>> +	if (ret == 0) {
>> +		gem_sync(fd, handles[0]);
> Not required for this test. However... You probably want to do the
> gem_sync() first. (Yes, there is an amusing reason to do :)

What reason is that and what do you mean by "first"?

>> +	for (try = 0, idx = 0; try < max_tries;) {
>> +		eb_handles[0] = handles[0];
>> +		eb_handles[1] = loc_handles[idx];
>> +		eb_handles[2] = loc_handles[idx + 1];
>> +
>> +		igt_assert(exec(fd, eb_handles, (uint32_t[2]){0, 0},
>> +				offsets) == 0);
>> +
>> +		if (offsets[1] > offsets[0]) {
>> +			distance = offsets[1] - offsets[0];
>> +			if (distance == PAGE_SIZE)
>> +				neighbours = true;
>> +			pad_to_size[0] = ALIGN(distance + PAGE_SIZE, PAGE_SIZE);
>> +		} else {
>> +			distance = offsets[0] - offsets[1];
>> +			if (distance == PAGE_SIZE)
>> +				neighbours = true;
>> +			pad_to_size[1] = ALIGN(distance + PAGE_SIZE, PAGE_SIZE);
>> +		}
>> +
>> +		if (neighbours)
>> +			break;
>> +
>> +		try++;
>> +		idx +=2;
>
> Just use idx++ here and allocate a new handle one at a time. Just as
> likely to be adjacent to the previous handle as the next one will be to

Ah yes, didn't think of that!

> us. For extra paranoia, you could even try an evict-everything pass :)

You mean if the lightweight approach fails? Ok.

Regards,

Tvrtko
On Wed, Apr 01, 2015 at 02:36:29PM +0100, Tvrtko Ursulin wrote:
> 
> On 04/01/2015 02:06 PM, Chris Wilson wrote:
> >On Wed, Apr 01, 2015 at 12:21:14PM +0100, Tvrtko Ursulin wrote:
> >>+	if (drmIoctl(fd, DRM_IOCTL_I915_GEM_EXECBUFFER2, &execbuf))
> >>+			ret = -errno;
> >
> >>+	if (ret == 0) {
> >>+		gem_sync(fd, handles[0]);
> >Not required for this test. However... You probably want to do the
> >gem_sync() first. (Yes, there is an amusing reason to do :)
> 
> What reason is that and what do you mean by "first"?

When calling the test multiple times successive passes will have a busy
batch, which could conceivably cause the relocations to fail, and the
returned offsets to be -1. Calling gem_sync() before execbuf prevents that
slowpath from affecting the test, without people wondering whether
you need to call gem_sync() before the returned array is valid (which is
why I did a double take here trying to work out what you meant the
gem_sync() to do).
 
> >Just use idx++ here and allocate a new handle one at a time. Just as
> >likely to be adjacent to the previous handle as the next one will be to
> 
> Ah yes, didn't think of that!
> 
> >us. For extra paranoia, you could even try an evict-everything pass :)
> 
> You mean if the lightweight approach fails? Ok.

Calling igt_drop_caches_set() is reasonable enough to in the preamble.
-Chris
On 04/01/2015 02:56 PM, Chris Wilson wrote:
> On Wed, Apr 01, 2015 at 02:36:29PM +0100, Tvrtko Ursulin wrote:
>>
>> On 04/01/2015 02:06 PM, Chris Wilson wrote:
>>> On Wed, Apr 01, 2015 at 12:21:14PM +0100, Tvrtko Ursulin wrote:
>>>> +	if (drmIoctl(fd, DRM_IOCTL_I915_GEM_EXECBUFFER2, &execbuf))
>>>> +			ret = -errno;
>>>
>>>> +	if (ret == 0) {
>>>> +		gem_sync(fd, handles[0]);
>>> Not required for this test. However... You probably want to do the
>>> gem_sync() first. (Yes, there is an amusing reason to do :)
>>
>> What reason is that and what do you mean by "first"?
>
> When calling the test multiple times successive passes will have a busy
> batch, which could conceivably cause the relocations to fail, and the
> returned offsets to be -1. Calling gem_sync() before execbuf prevents that
> slowpath from affecting the test, without people wondering whether
> you need to call gem_sync() before the returned array is valid (which is
> why I did a double take here trying to work out what you meant the
> gem_sync() to do).

How it may be busy if gem_sync after execbuf is supposed to wait until 
batch has been retired?

Regards,

Tvrtko
On Wed, Apr 01, 2015 at 03:14:26PM +0100, Tvrtko Ursulin wrote:
> 
> On 04/01/2015 02:56 PM, Chris Wilson wrote:
> >On Wed, Apr 01, 2015 at 02:36:29PM +0100, Tvrtko Ursulin wrote:
> >>
> >>On 04/01/2015 02:06 PM, Chris Wilson wrote:
> >>>On Wed, Apr 01, 2015 at 12:21:14PM +0100, Tvrtko Ursulin wrote:
> >>>>+	if (drmIoctl(fd, DRM_IOCTL_I915_GEM_EXECBUFFER2, &execbuf))
> >>>>+			ret = -errno;
> >>>
> >>>>+	if (ret == 0) {
> >>>>+		gem_sync(fd, handles[0]);
> >>>Not required for this test. However... You probably want to do the
> >>>gem_sync() first. (Yes, there is an amusing reason to do :)
> >>
> >>What reason is that and what do you mean by "first"?
> >
> >When calling the test multiple times successive passes will have a busy
> >batch, which could conceivably cause the relocations to fail, and the
> >returned offsets to be -1. Calling gem_sync() before execbuf prevents that
> >slowpath from affecting the test, without people wondering whether
> >you need to call gem_sync() before the returned array is valid (which is
> >why I did a double take here trying to work out what you meant the
> >gem_sync() to do).
> 
> How it may be busy if gem_sync after execbuf is supposed to wait
> until batch has been retired?

It's not. I am trying to say that the placement of gem_sync() here
suggests to me that you thought it was required to get valid offsets
after an execbuf call. But in actualilty you do not need a gem_sync() in
this test at all as you are not doing relocations.
-Chris