igt/gem_close_race: Trim object size

The challenge here is to race GPU activity versus the gem_close. Using a
large objects makes the window of GPU activity larger - except on !llc
systems we then incur massive overhead from clflush, likely destroying
any race (due to mutex contention). And that mutex contention is
amplified by the number of cores - bad news for Baytrail.

Give up and make the objects smaller and hope that the test is run
frequently enough to catch the race.

References: https://bugs.freedesktop.org/show_bug.cgi?id=71029
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

1 files changed, 5 insertions, 5 deletions

diff --git a/tests/gem_close_race.c b/tests/gem_close_race.c
index e46e2c15..02a408a9 100644
--- a/tests/gem_close_race.c
+++ b/tests/gem_close_race.c
@@ -42,7 +42,7 @@
 #include "drmtest.h"
 #include "intel_chipset.h"
 
-#define OBJECT_SIZE 1024*1024*4
+#define OBJECT_SIZE 4096
 
 #define COPY_BLT_CMD		(2<<29|0x53<<22|0x6)
 #define BLT_WRITE_ALPHA		(1<<21)
@@ -71,7 +71,7 @@ static void selfcopy(int fd, uint32_t handle, int loops)
 	b++;
 	*b++ = 0xcc << 16 | 1 << 25 | 1 << 24 | (4*1024);
 	*b++ = 0;
-	*b++ = 512 << 16 | 1024;
+	*b++ = 1 << 16 | 1024;
 
 	reloc[0].offset = (b - buf) * sizeof(*b);
 	reloc[0].target_handle = handle;
@@ -130,7 +130,7 @@ static uint32_t load(int fd)
 	if (handle == 0)
 		return 0;
 
-	selfcopy(fd, handle, 30);
+	selfcopy(fd, handle, 100);
 	return handle;
 }
 
@@ -172,7 +172,7 @@ static void *thread_run(void *_data)
 			if (create.handle == 0)
 				continue;
 
-			selfcopy(t->fds[n], create.handle, 10);
+			selfcopy(t->fds[n], create.handle, 100);
 
 			drmIoctl(t->fds[n], DRM_IOCTL_GEM_CLOSE, &create.handle);
 		}
@@ -204,7 +204,7 @@ static void *thread_busy(void *_data)
 		if (create.handle == 0)
 			continue;
 
-		selfcopy(t->fds[n], create.handle, 1);
+		selfcopy(t->fds[n], create.handle, 10);
 
 		busy.handle = create.handle;
 		drmIoctl(t->fds[n], DRM_IOCTL_I915_GEM_BUSY, &busy);