intel-gfx.lists.freedesktop.org archive mirror
 help / color / mirror / Atom feed
From: Chris Wilson <chris@chris-wilson.co.uk>
To: intel-gfx@lists.freedesktop.org
Cc: igt-dev@lists.freedesktop.org
Subject: [PATCH i-g-t 02/12] i915/gem_ctx_param: Test set/get (copy) VM
Date: Wed, 22 May 2019 12:37:02 +0100	[thread overview]
Message-ID: <20190522113712.23512-2-chris@chris-wilson.co.uk> (raw)
In-Reply-To: <20190522113712.23512-1-chris@chris-wilson.co.uk>

Exercise reusing the GTT of one ctx in another.

v2: Test setting back to the same VM
v3: Check the VM still exists after the parent ctx are dead.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
---
 tests/i915/gem_ctx_param.c | 116 +++++++++++++++++++++++++++++++++----
 1 file changed, 104 insertions(+), 12 deletions(-)

diff --git a/tests/i915/gem_ctx_param.c b/tests/i915/gem_ctx_param.c
index b6f57236c..fa0ab5202 100644
--- a/tests/i915/gem_ctx_param.c
+++ b/tests/i915/gem_ctx_param.c
@@ -28,6 +28,7 @@
 #include <limits.h>
 
 #include "igt.h"
+#include "i915/gem_vm.h"
 
 IGT_TEST_DESCRIPTION("Basic test for context set/get param input validation.");
 
@@ -36,17 +37,6 @@ IGT_TEST_DESCRIPTION("Basic test for context set/get param input validation.");
 #define NEW_CTX	BIT(0)
 #define USER BIT(1)
 
-static int reopen_driver(int fd)
-{
-	char path[256];
-
-	snprintf(path, sizeof(path), "/proc/self/fd/%d", fd);
-	fd = open(path, O_RDWR);
-	igt_assert_lte(0, fd);
-
-	return fd;
-}
-
 static void set_priority(int i915)
 {
 	static const int64_t test_values[] = {
@@ -91,7 +81,7 @@ static void set_priority(int i915)
 	igt_permute_array(values, size, igt_exchange_int64);
 
 	igt_fork(flags, NEW_CTX | USER) {
-		int fd = reopen_driver(i915);
+		int fd = gem_reopen_driver(i915);
 		struct drm_i915_gem_context_param arg = {
 			.param = I915_CONTEXT_PARAM_PRIORITY,
 			.ctx_id = flags & NEW_CTX ? gem_context_create(fd) : 0,
@@ -143,6 +133,105 @@ static void set_priority(int i915)
 	free(values);
 }
 
+static uint32_t __batch_create(int i915, uint32_t offset)
+{
+	const uint32_t bbe = MI_BATCH_BUFFER_END;
+	uint32_t handle;
+
+	handle = gem_create(i915, ALIGN(offset + 4, 4096));
+	gem_write(i915, handle, offset, &bbe, sizeof(bbe));
+
+	return handle;
+}
+
+static uint32_t batch_create(int i915)
+{
+	return __batch_create(i915, 0);
+}
+
+static void test_vm(int i915)
+{
+	const uint64_t nonzero_offset = 48 << 20;
+	struct drm_i915_gem_exec_object2 batch = {
+		.handle = batch_create(i915),
+	};
+	struct drm_i915_gem_execbuffer2 eb = {
+		.buffers_ptr = to_user_pointer(&batch),
+		.buffer_count = 1,
+	};
+	struct drm_i915_gem_context_param arg = {
+		.param = I915_CONTEXT_PARAM_VM,
+	};
+	uint32_t parent, child;
+
+	/*
+	 * Proving 2 contexts share the same GTT is quite tricky as we have no
+	 * means of directly comparing them (each handle returned to userspace
+	 * is unique). What we do instead is rely on a quirk of execbuf that
+	 * it does not try to move an VMA without good reason, and so that
+	 * having used an object in one context, it will have the same address
+	 * in the next context that shared the VM.
+	 */
+
+	arg.value = -1ull;
+	igt_require(__gem_context_set_param(i915, &arg) == -ENOENT);
+
+	parent = gem_context_create(i915);
+	child = gem_context_create(i915);
+
+	/* Using implicit soft-pinning */
+	eb.rsvd1 = parent;
+	batch.offset = nonzero_offset;
+	gem_execbuf(i915, &eb);
+	igt_assert_eq_u64(batch.offset, nonzero_offset);
+
+	eb.rsvd1 = child;
+	batch.offset = 0;
+	gem_execbuf(i915, &eb);
+	igt_assert_eq_u64(batch.offset, 0);
+
+	eb.rsvd1 = parent;
+	gem_execbuf(i915, &eb);
+	igt_assert_eq_u64(batch.offset, nonzero_offset);
+
+	arg.ctx_id = parent;
+	gem_context_get_param(i915, &arg);
+	gem_context_set_param(i915, &arg);
+
+	/* Still the same VM, so expect the old VMA again */
+	batch.offset = 0;
+	gem_execbuf(i915, &eb);
+	igt_assert_eq_u64(batch.offset, nonzero_offset);
+
+	arg.ctx_id = child;
+	gem_context_set_param(i915, &arg);
+
+	eb.rsvd1 = child;
+	batch.offset = 0;
+	gem_execbuf(i915, &eb);
+	igt_assert_eq_u64(batch.offset, nonzero_offset);
+
+	gem_context_destroy(i915, child);
+	gem_context_destroy(i915, parent);
+
+	/* both contexts destroyed, but we still keep hold of the vm */
+	child = gem_context_create(i915);
+
+	arg.ctx_id = child;
+	gem_context_set_param(i915, &arg);
+
+	eb.rsvd1 = child;
+	batch.offset = 0;
+	gem_execbuf(i915, &eb);
+	igt_assert_eq_u64(batch.offset, nonzero_offset);
+
+	gem_context_destroy(i915, child);
+	gem_vm_destroy(i915, arg.value);
+
+	gem_sync(i915, batch.handle);
+	gem_close(i915, batch.handle);
+}
+
 igt_main
 {
 	struct drm_i915_gem_context_param arg;
@@ -253,6 +342,9 @@ igt_main
 		gem_context_set_param(fd, &arg);
 	}
 
+	igt_subtest("vm")
+		test_vm(fd);
+
 	arg.param = I915_CONTEXT_PARAM_PRIORITY;
 
 	igt_subtest("set-priority-not-supported") {
-- 
2.20.1

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

  reply	other threads:[~2019-05-22 11:37 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-05-22 11:37 [PATCH i-g-t 01/12] i915: Add gem_vm_create Chris Wilson
2019-05-22 11:37 ` Chris Wilson [this message]
2019-05-22 11:37 ` [PATCH i-g-t 03/12] i915/gem_ctx_create: Basic checks for constructor properties Chris Wilson
2019-05-23 13:41   ` Tvrtko Ursulin
2019-05-22 11:37 ` [PATCH i-g-t 04/12] i915: Add gem_ctx_clone Chris Wilson
2019-05-22 11:37 ` [PATCH i-g-t 05/12] i915: Exercise creating context with shared GTT Chris Wilson
2019-05-23 13:39   ` Tvrtko Ursulin
2019-05-22 11:37 ` [PATCH i-g-t 06/12] i915/gem_ctx_switch: Exercise queues Chris Wilson
2019-05-22 11:37 ` [PATCH i-g-t 07/12] i915/gem_exec_whisper: Fork all-engine tests one-per-engine Chris Wilson
2019-05-22 11:37 ` [PATCH i-g-t 08/12] i915: Add gem_ctx_engines Chris Wilson
2019-05-22 11:37 ` [PATCH i-g-t 09/12] i915: Add gem_exec_balancer Chris Wilson
2019-05-22 12:13   ` Tvrtko Ursulin
2019-05-22 11:37 ` [PATCH i-g-t 10/12] i915/gem_exec_balancer: Exercise bonded pairs Chris Wilson
2019-05-22 12:26   ` Tvrtko Ursulin
2019-05-22 12:33     ` Chris Wilson
2019-05-22 12:52   ` [PATCH i-g-t] " Chris Wilson
2019-05-22 13:08     ` Tvrtko Ursulin
2019-05-22 11:37 ` [PATCH i-g-t 11/12] i915/gem_exec_latency: Measure the latency of context switching Chris Wilson
2019-05-22 11:37 ` [PATCH i-g-t 12/12] i915/gem_exec_latency: Add another variant of waiter latency Chris Wilson

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=20190522113712.23512-2-chris@chris-wilson.co.uk \
    --to=chris@chris-wilson.co.uk \
    --cc=igt-dev@lists.freedesktop.org \
    --cc=intel-gfx@lists.freedesktop.org \
    /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;
as well as URLs for NNTP newsgroup(s).