* [PATCH] tests/drv_hangman: skip a few asserts when using the cmd parser
@ 2014-11-04 22:00 bradley.d.volkin
2014-11-05 12:36 ` Daniel Vetter
0 siblings, 1 reply; 2+ messages in thread
From: bradley.d.volkin @ 2014-11-04 22:00 UTC (permalink / raw)
To: intel-gfx
From: Brad Volkin <bradley.d.volkin@intel.com>
This test has a few checks that batch buffer addresses in the error
state match the expected address for the userspace supplied batch.
But the batch buffer copy piece of the command parser means that
the logged addresses are actually _supposed_ to be different. So
skip just those checks.
Signed-off-by: Brad Volkin <bradley.d.volkin@intel.com>
---
tests/drv_hangman.c | 43 +++++++++++++++++++++++++++++++++++++------
1 file changed, 37 insertions(+), 6 deletions(-)
I'm not sure it's actually worth doing any of the work beyond the
'if (bb_matched == 1)' block in check_error_state when the command
parser is enabled. Here I've kept that work and just taken out the
one check that would fail. But it seems like the work that's left
is only there to enable the removed check, so maybe not needed.
I'll defer to those that know more about this test.
diff --git a/tests/drv_hangman.c b/tests/drv_hangman.c
index 3d6b98b..8fbc2d3 100644
--- a/tests/drv_hangman.c
+++ b/tests/drv_hangman.c
@@ -36,6 +36,10 @@
#include "igt_debugfs.h"
#include "ioctl_wrappers.h"
+#ifndef I915_PARAM_CMD_PARSER_VERSION
+#define I915_PARAM_CMD_PARSER_VERSION 28
+#endif
+
static int _read_sysfs(void *dst, int maxlen,
const char* path,
const char *fname)
@@ -262,6 +266,7 @@ static void test_error_state_basic(void)
}
static void check_error_state(const int gen,
+ const bool uses_cmd_parser,
const char *expected_ring_name,
uint64_t expected_offset)
{
@@ -300,7 +305,8 @@ static void check_error_state(const int gen,
char expected_line[32];
igt_assert(strstr(ring_name, expected_ring_name));
- igt_assert(gtt_offset == expected_offset);
+ if (!uses_cmd_parser)
+ igt_assert(gtt_offset == expected_offset);
for (i = 0; i < sizeof(batch) / 4; i++) {
igt_assert(getline(&line, &line_size, file) > 0);
@@ -352,10 +358,12 @@ static void check_error_state(const int gen,
i++;
}
}
- if (gen >= 4)
- igt_assert(expected_addr == expected_offset);
- else
- igt_assert((expected_addr & ~0x1) == expected_offset);
+ if (!uses_cmd_parser) {
+ if (gen >= 4)
+ igt_assert(expected_addr == expected_offset);
+ else
+ igt_assert((expected_addr & ~0x1) == expected_offset);
+ }
ringbuf_ok = true;
continue;
}
@@ -370,22 +378,45 @@ static void check_error_state(const int gen,
close(debug_fd);
}
+static bool uses_cmd_parser(int fd, int gen)
+{
+ int parser_version = 0;
+ drm_i915_getparam_t gp;
+ int rc;
+
+ gp.param = I915_PARAM_CMD_PARSER_VERSION;
+ gp.value = &parser_version;
+ rc = drmIoctl(fd, DRM_IOCTL_I915_GETPARAM, &gp);
+ if (rc || parser_version == 0)
+ return false;
+
+ if (!gem_uses_aliasing_ppgtt(fd))
+ return false;
+
+ if (gen != 7)
+ return false;
+
+ return true;
+}
+
static void test_error_state_capture(unsigned ring_id,
const char *ring_name)
{
int fd, gen;
uint64_t offset;
+ bool cmd_parser;
check_other_clients();
clear_error_state();
fd = drm_open_any();
gen = intel_gen(intel_get_drm_devid(fd));
+ cmd_parser = uses_cmd_parser(fd, gen);
offset = submit_batch(fd, ring_id, true);
close(fd);
- check_error_state(gen, ring_name, offset);
+ check_error_state(gen, cmd_parser, ring_name, offset);
}
static const struct target_ring {
--
1.9.1
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply related [flat|nested] 2+ messages in thread
* Re: [PATCH] tests/drv_hangman: skip a few asserts when using the cmd parser
2014-11-04 22:00 [PATCH] tests/drv_hangman: skip a few asserts when using the cmd parser bradley.d.volkin
@ 2014-11-05 12:36 ` Daniel Vetter
0 siblings, 0 replies; 2+ messages in thread
From: Daniel Vetter @ 2014-11-05 12:36 UTC (permalink / raw)
To: bradley.d.volkin; +Cc: intel-gfx
On Tue, Nov 04, 2014 at 02:00:43PM -0800, bradley.d.volkin@intel.com wrote:
> +static bool uses_cmd_parser(int fd, int gen)
> +{
> + int parser_version = 0;
> + drm_i915_getparam_t gp;
> + int rc;
> +
> + gp.param = I915_PARAM_CMD_PARSER_VERSION;
> + gp.value = &parser_version;
> + rc = drmIoctl(fd, DRM_IOCTL_I915_GETPARAM, &gp);
> + if (rc || parser_version == 0)
> + return false;
> +
> + if (!gem_uses_aliasing_ppgtt(fd))
> + return false;
> +
> + if (gen != 7)
> + return false;
Hm, these additional checks are a bit underwhelming, but I guess so be it.
Since we have the cmd parser only on gen7 test coverage isn't really
seriously reduced anyway.
Applied, thanks for the patch.
-Daniel
--
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2014-11-05 12:36 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2014-11-04 22:00 [PATCH] tests/drv_hangman: skip a few asserts when using the cmd parser bradley.d.volkin
2014-11-05 12:36 ` Daniel Vetter
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox