ntel-gpu-tools/tests/kms_flip.c
Daniel Vetter 304c458d54 tests/kms_flip: Skip if no clone configuration could be found
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-10-07 22:55:14 +02:00

1506 lines
40 KiB
C

/*
* Copyright 2012 Intel Corporation
* Jesse Barnes <jesse.barnes@intel.com>
*
* 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 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.
*/
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#include <cairo.h>
#include <errno.h>
#include <fcntl.h>
#include <math.h>
#include <stdint.h>
#include <unistd.h>
#include <sys/poll.h>
#include <sys/time.h>
#include <sys/ioctl.h>
#include <linux/kd.h>
#include "i915_drm.h"
#include "drmtest.h"
#include "testdisplay.h"
#include "rendercopy.h"
#define TEST_DPMS (1 << 0)
#define TEST_WITH_DUMMY_BCS (1 << 1)
#define TEST_WITH_DUMMY_RCS (1 << 2)
#define TEST_PAN (1 << 3)
#define TEST_MODESET (1 << 4)
#define TEST_CHECK_TS (1 << 5)
#define TEST_EBUSY (1 << 6)
#define TEST_EINVAL (1 << 7)
#define TEST_FLIP (1 << 8)
#define TEST_VBLANK (1 << 9)
#define TEST_VBLANK_BLOCK (1 << 10)
#define TEST_VBLANK_ABSOLUTE (1 << 11)
#define TEST_VBLANK_EXPIRED_SEQ (1 << 12)
#define TEST_FB_RECREATE (1 << 13)
#define TEST_RMFB (1 << 14)
#define TEST_HANG (1 << 15)
#define TEST_NOEVENT (1 << 16)
#define TEST_FB_BAD_TILING (1 << 17)
#define TEST_SINGLE_BUFFER (1 << 18)
#define TEST_DPMS_OFF (1 << 19)
#define TEST_NO_2X_OUTPUT (1 << 20)
#define TEST_DPMS_OFF_OTHERS (1 << 21)
#define EVENT_FLIP (1 << 0)
#define EVENT_VBLANK (1 << 1)
#ifndef DRM_CAP_TIMESTAMP_MONOTONIC
#define DRM_CAP_TIMESTAMP_MONOTONIC 6
#endif
drmModeRes *resources;
int drm_fd;
static drm_intel_bufmgr *bufmgr;
struct intel_batchbuffer *batch;
uint32_t devid;
int test_time = 3;
static bool monotonic_timestamp;
static drmModeConnector *last_connector;
uint32_t *fb_ptr;
struct type_name {
int type;
const char *name;
};
struct event_state {
const char *name;
/*
* Event data for the last event that has already passed our check.
* Updated using the below current_* vars in update_state().
*/
struct timeval last_ts; /* kernel reported timestamp */
struct timeval last_received_ts; /* the moment we received it */
unsigned int last_seq; /* kernel reported seq. num */
/*
* Event data for for the current event that we just received and
* going to check for validity. Set in event_handler().
*/
struct timeval current_ts; /* kernel reported timestamp */
struct timeval current_received_ts; /* the moment we received it */
unsigned int current_seq; /* kernel reported seq. num */
int count; /* # of events of this type */
/* Step between the current and next 'target' sequence number. */
int seq_step;
};
struct test_output {
int mode_valid;
drmModeModeInfo kmode[4];
drmModeEncoder *kencoder[4];
drmModeConnector *kconnector[4];
uint32_t _connector[4];
uint32_t _crtc[4];
int count; /* 1:1 mapping between crtc:connector */
int flags;
int pipe; /* primary pipe for vblank */
unsigned int current_fb_id;
unsigned int fb_width;
unsigned int fb_height;
unsigned int fb_ids[3];
int bpp, depth;
struct kmstest_fb fb_info[3];
struct event_state flip_state;
struct event_state vblank_state;
unsigned int pending_events;
int flip_count;
};
static unsigned long gettime_us(void)
{
struct timespec ts;
clock_gettime(CLOCK_MONOTONIC, &ts);
return ts.tv_sec * 1000000 + ts.tv_nsec / 1000;
}
static void emit_dummy_load__bcs(struct test_output *o)
{
int i, limit;
drm_intel_bo *dummy_bo, *target_bo, *tmp_bo;
struct kmstest_fb *fb_info = &o->fb_info[o->current_fb_id];
unsigned pitch = fb_info->stride;
limit = intel_gen(devid) < 6 ? 500 : 5000;
dummy_bo = drm_intel_bo_alloc(bufmgr, "dummy_bo", fb_info->size, 4096);
igt_assert(dummy_bo);
target_bo = gem_handle_to_libdrm_bo(bufmgr, drm_fd, "imported", fb_info->gem_handle);
igt_assert(target_bo);
for (i = 0; i < limit; i++) {
BEGIN_BATCH(8);
OUT_BATCH(XY_SRC_COPY_BLT_CMD |
XY_SRC_COPY_BLT_WRITE_ALPHA |
XY_SRC_COPY_BLT_WRITE_RGB);
OUT_BATCH((3 << 24) | /* 32 bits */
(0xcc << 16) | /* copy ROP */
pitch);
OUT_BATCH(0 << 16 | 0);
OUT_BATCH(o->fb_height << 16 | o->fb_width);
OUT_RELOC_FENCED(dummy_bo, I915_GEM_DOMAIN_RENDER, I915_GEM_DOMAIN_RENDER, 0);
OUT_BATCH(0 << 16 | 0);
OUT_BATCH(pitch);
OUT_RELOC_FENCED(target_bo, I915_GEM_DOMAIN_RENDER, 0, 0);
ADVANCE_BATCH();
if (IS_GEN6(devid) || IS_GEN7(devid)) {
BEGIN_BATCH(3);
OUT_BATCH(XY_SETUP_CLIP_BLT_CMD);
OUT_BATCH(0);
OUT_BATCH(0);
ADVANCE_BATCH();
}
tmp_bo = dummy_bo;
dummy_bo = target_bo;
target_bo = tmp_bo;
}
intel_batchbuffer_flush(batch);
drm_intel_bo_unreference(dummy_bo);
drm_intel_bo_unreference(target_bo);
}
static void emit_dummy_load__rcs(struct test_output *o)
{
const struct kmstest_fb *fb_info = &o->fb_info[o->current_fb_id];
render_copyfunc_t copyfunc;
struct scratch_buf sb[2], *src, *dst;
int i, limit;
copyfunc = get_render_copyfunc(devid);
if (copyfunc == NULL)
return emit_dummy_load__bcs(o);
limit = intel_gen(devid) < 6 ? 500 : 5000;
sb[0].bo = drm_intel_bo_alloc(bufmgr, "dummy_bo", fb_info->size, 4096);
igt_assert(sb[0].bo);
sb[0].size = sb[0].bo->size;
sb[0].tiling = I915_TILING_NONE;
sb[0].data = NULL;
sb[0].num_tiles = sb[0].bo->size;
sb[0].stride = 4 * o->fb_width;
sb[1].bo = gem_handle_to_libdrm_bo(bufmgr, drm_fd, "imported", fb_info->gem_handle);
igt_assert(sb[1].bo);
sb[1].size = sb[1].bo->size;
sb[1].tiling = fb_info->tiling;
sb[1].data = NULL;
sb[1].num_tiles = sb[1].bo->size;
sb[1].stride = fb_info->stride;
src = &sb[0];
dst = &sb[1];
for (i = 0; i < limit; i++) {
struct scratch_buf *tmp;
copyfunc(batch,
src, 0, 0,
o->fb_width, o->fb_height,
dst, 0, 0);
tmp = src;
src = dst;
dst = tmp;
}
intel_batchbuffer_flush(batch);
drm_intel_bo_unreference(sb[0].bo);
drm_intel_bo_unreference(sb[1].bo);
}
static void set_connector_dpms(drmModeConnector *connector, int mode)
{
int i, dpms = 0;
bool found_it = false;
for (i = 0; i < connector->count_props; i++) {
struct drm_mode_get_property prop;
prop.prop_id = connector->props[i];
prop.count_values = 0;
prop.count_enum_blobs = 0;
if (drmIoctl(drm_fd, DRM_IOCTL_MODE_GETPROPERTY, &prop))
continue;
if (strcmp(prop.name, "DPMS"))
continue;
dpms = prop.prop_id;
found_it = true;
break;
}
igt_assert_f(found_it, "DPMS property not found on %d\n",
connector->connector_id);
igt_assert(drmModeConnectorSetProperty(drm_fd, connector->connector_id,
dpms, mode) == 0);
}
static void dpms_off_other_outputs(struct test_output *o)
{
int i, n;
drmModeConnector *connector;
uint32_t connector_id;
for (i = 0; i < resources->count_connectors; i++) {
connector_id = resources->connectors[i];
for (n = 0; n < o->count; n++) {
if (connector_id == o->kconnector[n]->connector_id)
goto next;
}
connector = drmModeGetConnector(drm_fd, connector_id);
set_connector_dpms(connector, DRM_MODE_DPMS_ON);
set_connector_dpms(connector, DRM_MODE_DPMS_OFF);
next:
;
}
}
static void set_dpms(struct test_output *o, int mode)
{
for (int n = 0; n < o->count; n++)
set_connector_dpms(o->kconnector[n], mode);
}
static void set_flag(unsigned int *v, unsigned int flag)
{
igt_assert(!(*v & flag));
*v |= flag;
}
static void clear_flag(unsigned int *v, unsigned int flag)
{
igt_assert(*v & flag);
*v &= ~flag;
}
static int do_page_flip(struct test_output *o, int fb_id, bool event)
{
int n, ret = 0;
o->flip_count = 0;
for (n = 0; ret == 0 && n < o->count; n++)
ret = drmModePageFlip(drm_fd, o->_crtc[n], fb_id,
event ? DRM_MODE_PAGE_FLIP_EVENT : 0,
event ? (void *)((unsigned long)o | (n==0)) : NULL);
if (ret == 0 && event)
set_flag(&o->pending_events, EVENT_FLIP);
return ret;
}
struct vblank_reply {
unsigned int sequence;
struct timeval ts;
};
static int __wait_for_vblank(unsigned int flags, int crtc_idx,
int target_seq, unsigned long ret_data,
struct vblank_reply *reply)
{
drmVBlank wait_vbl;
int ret;
unsigned crtc_idx_mask;
bool event = !(flags & TEST_VBLANK_BLOCK);
memset(&wait_vbl, 0, sizeof(wait_vbl));
crtc_idx_mask = crtc_idx << DRM_VBLANK_HIGH_CRTC_SHIFT;
igt_assert(!(crtc_idx_mask & ~DRM_VBLANK_HIGH_CRTC_MASK));
wait_vbl.request.type = crtc_idx_mask;
if (flags & TEST_VBLANK_ABSOLUTE)
wait_vbl.request.type |= DRM_VBLANK_ABSOLUTE;
else
wait_vbl.request.type |= DRM_VBLANK_RELATIVE;
if (event) {
wait_vbl.request.type |= DRM_VBLANK_EVENT;
wait_vbl.request.signal = ret_data;
}
wait_vbl.request.sequence = target_seq;
ret = drmWaitVBlank(drm_fd, &wait_vbl);
if (ret == 0) {
reply->ts.tv_sec = wait_vbl.reply.tval_sec;
reply->ts.tv_usec = wait_vbl.reply.tval_usec;
reply->sequence = wait_vbl.reply.sequence;
} else
ret = -errno;
return ret;
}
static int do_wait_for_vblank(struct test_output *o, int pipe_id,
int target_seq, struct vblank_reply *reply)
{
int ret;
ret = __wait_for_vblank(o->flags, pipe_id, target_seq, (unsigned long)o,
reply);
if (ret == 0 && !(o->flags & TEST_VBLANK_BLOCK))
set_flag(&o->pending_events, EVENT_VBLANK);
return ret;
}
static bool
analog_tv_connector(struct test_output *o)
{
uint32_t connector_type = o->kconnector[0]->connector_type;
return connector_type == DRM_MODE_CONNECTOR_TV ||
connector_type == DRM_MODE_CONNECTOR_9PinDIN ||
connector_type == DRM_MODE_CONNECTOR_SVIDEO ||
connector_type == DRM_MODE_CONNECTOR_Composite;
}
static void event_handler(struct event_state *es, unsigned int frame,
unsigned int sec, unsigned int usec)
{
struct timeval now;
if (monotonic_timestamp) {
struct timespec ts;
clock_gettime(CLOCK_MONOTONIC, &ts);
now.tv_sec = ts.tv_sec;
now.tv_usec = ts.tv_nsec / 1000;
} else {
gettimeofday(&now, NULL);
}
es->current_received_ts = now;
es->current_ts.tv_sec = sec;
es->current_ts.tv_usec = usec;
es->current_seq = frame;
}
static void page_flip_handler(int fd, unsigned int frame, unsigned int sec,
unsigned int usec, void *data)
{
int primary = (unsigned long)data & 1;
struct test_output *o = (void *)((unsigned long)data & ~ 1);
if (++o->flip_count == o->count)
clear_flag(&o->pending_events, EVENT_FLIP);
if (primary)
event_handler(&o->flip_state, frame, sec, usec);
}
static double frame_time(struct test_output *o)
{
return 1000.0 * 1000.0 / o->kmode[0].vrefresh;
}
static void fixup_premature_vblank_ts(struct test_output *o,
struct event_state *es)
{
/*
* In case a power off event preempts the completion of a
* wait-for-vblank event the kernel will return a wf-vblank event with
* a zeroed-out timestamp. In order that check_state() doesn't
* complain replace this ts with a valid ts. As we can't calculate the
* exact timestamp, just use the time we received the event.
*/
struct timeval tv;
if (!(o->flags & (TEST_DPMS | TEST_MODESET)))
return;
if (o->vblank_state.current_ts.tv_sec != 0 ||
o->vblank_state.current_ts.tv_usec != 0)
return;
tv.tv_sec = 0;
tv.tv_usec = 1;
timersub(&es->current_received_ts, &tv, &es->current_ts);
}
static void vblank_handler(int fd, unsigned int frame, unsigned int sec,
unsigned int usec, void *data)
{
struct test_output *o = data;
clear_flag(&o->pending_events, EVENT_VBLANK);
event_handler(&o->vblank_state, frame, sec, usec);
fixup_premature_vblank_ts(o, &o->vblank_state);
}
static void check_state(struct test_output *o, struct event_state *es)
{
struct timeval diff;
double usec_interflip;
timersub(&es->current_ts, &es->current_received_ts, &diff);
if ((!analog_tv_connector(o)) &&
(diff.tv_sec > 0 || (diff.tv_sec == 0 && diff.tv_usec > 2000))) {
fprintf(stderr, "%s ts delayed for too long: %is, %iusec\n",
es->name, (int)diff.tv_sec, (int)diff.tv_usec);
igt_fail(5);
}
if (es->count == 0)
return;
if (!timercmp(&es->last_received_ts, &es->current_ts, <)) {
fprintf(stderr, "%s ts before the %s was issued!\n",
es->name, es->name);
timersub(&es->current_ts, &es->last_received_ts, &diff);
fprintf(stderr, "timerdiff %is, %ius\n",
(int) diff.tv_sec, (int) diff.tv_usec);
igt_fail(6);
}
/* This bounding matches the one in DRM_IOCTL_WAIT_VBLANK. */
if (!(o->flags & (TEST_DPMS | TEST_MODESET))) {
/* check only valid if no modeset happens in between, that
* increments by (1 << 23) on each step. */
if (es->current_seq - (es->last_seq + es->seq_step) > 1UL << 23) {
fprintf(stderr, "unexpected %s seq %u, should be >= %u\n",
es->name, es->current_seq, es->last_seq + es->seq_step);
igt_fail(10);
}
}
if ((o->flags & TEST_CHECK_TS) && (!analog_tv_connector(o))) {
timersub(&es->current_ts, &es->last_ts, &diff);
usec_interflip = (double)es->seq_step * frame_time(o);
if (fabs((((double) diff.tv_usec) - usec_interflip) /
usec_interflip) > 0.005) {
fprintf(stderr, "inter-%s ts jitter: %is, %ius\n",
es->name,
(int) diff.tv_sec, (int) diff.tv_usec);
igt_fail(9);
}
if (es->current_seq != es->last_seq + es->seq_step) {
fprintf(stderr, "unexpected %s seq %u, expected %u\n",
es->name, es->current_seq,
es->last_seq + es->seq_step);
igt_fail(9);
}
}
}
static void check_state_correlation(struct test_output *o,
struct event_state *es1,
struct event_state *es2)
{
struct timeval tv_diff;
double ftime;
double usec_diff;
int seq_diff;
if (es1->count == 0 || es2->count == 0)
return;
timersub(&es2->current_ts, &es1->current_ts, &tv_diff);
usec_diff = tv_diff.tv_sec * 1000 * 1000 + tv_diff.tv_usec;
seq_diff = es2->current_seq - es1->current_seq;
ftime = frame_time(o);
usec_diff -= seq_diff * ftime;
igt_assert_f(fabs(usec_diff) / ftime <= 0.005,
"timestamp mismatch between %s and %s (diff %.4f sec)\n",
es1->name, es2->name, usec_diff / 1000 / 1000);
}
static void check_all_state(struct test_output *o,
unsigned int completed_events)
{
bool flip, vblank;
flip = completed_events & EVENT_FLIP;
vblank = completed_events & EVENT_VBLANK;
if (flip)
check_state(o, &o->flip_state);
if (vblank)
check_state(o, &o->vblank_state);
if (flip && vblank)
check_state_correlation(o, &o->flip_state, &o->vblank_state);
}
static void recreate_fb(struct test_output *o)
{
drmModeFBPtr r;
struct kmstest_fb *fb_info = &o->fb_info[o->current_fb_id];
uint32_t new_fb_id;
/* Call rmfb/getfb/addfb to ensure those don't introduce stalls */
r = drmModeGetFB(drm_fd, fb_info->fb_id);
igt_assert(r);
do_or_die(drmModeAddFB(drm_fd, o->fb_width, o->fb_height, o->depth,
o->bpp, fb_info->stride,
r->handle, &new_fb_id));
gem_close(drm_fd, r->handle);
drmFree(r);
do_or_die(drmModeRmFB(drm_fd, fb_info->fb_id));
o->fb_ids[o->current_fb_id] = new_fb_id;
o->fb_info[o->current_fb_id].fb_id = new_fb_id;
}
static void set_y_tiling(struct test_output *o, int fb_idx)
{
drmModeFBPtr r;
struct kmstest_fb *fb_info = &o->fb_info[fb_idx];
/* Call rmfb/getfb/addfb to ensure those don't introduce stalls */
r = drmModeGetFB(drm_fd, fb_info->fb_id);
igt_assert(r);
gem_set_tiling(drm_fd, r->handle, I915_TILING_Y, fb_info->stride);
gem_close(drm_fd, r->handle);
drmFree(r);
}
static void exec_nop(int fd, uint32_t handle)
{
struct drm_i915_gem_execbuffer2 execbuf;
struct drm_i915_gem_exec_object2 gem_exec[1];
uint32_t b[2] = {MI_BATCH_BUFFER_END};
gem_write(fd, handle, 0, b, sizeof(b));
gem_exec[0].handle = handle;
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 = 0;
gem_exec[0].rsvd1 = 0;
gem_exec[0].rsvd2 = 0;
execbuf.buffers_ptr = (uintptr_t)gem_exec;
execbuf.buffer_count = 1;
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;
igt_assert(drmIoctl(fd,
DRM_IOCTL_I915_GEM_EXECBUFFER2,
&execbuf) == 0);
}
static void eat_error_state(struct test_output *o)
{
static const char dfs_base[] = "/sys/kernel/debug/dri";
static const char dfs_entry_error[] = "i915_error_state";
static const char dfs_entry_stop[] = "i915_ring_stop";
static const char data[] = "";
static char tmp[128];
char fname[FILENAME_MAX];
int card_index = drm_get_card();
int fd;
igt_assert(card_index != -1);
/* clear the error state */
snprintf(fname, FILENAME_MAX, "%s/%i/%s",
dfs_base, card_index, dfs_entry_error);
fd = open(fname, O_WRONLY);
igt_assert(fd >= 0);
igt_assert(write(fd, data, sizeof(data)) == sizeof(data));
close(fd);
/* and check whether stop_rings is not reset, i.e. the hang has indeed
* happened */
snprintf(fname, FILENAME_MAX, "%s/%i/%s",
dfs_base, card_index, dfs_entry_stop);
fd = open(fname, O_RDONLY);
igt_assert(fd >= 0);
igt_assert(read(fd, tmp, sizeof tmp) > 0);
igt_assert_f(atoi(tmp) == 0,
"no gpu hang detected, stop_rings is still %s\n", tmp);
close(fd);
}
static void hang_gpu(struct test_output *o)
{
static const char dfs_base[] = "/sys/kernel/debug/dri";
static const char dfs_entry[] = "i915_ring_stop";
static const char data[] = "0xf";
char fname[FILENAME_MAX];
int card_index = drm_get_card();
int fd;
snprintf(fname, FILENAME_MAX, "%s/%i/%s",
dfs_base, card_index, dfs_entry);
fd = open(fname, O_WRONLY);
igt_assert(fd >= 0);
igt_assert(write(fd, data, sizeof(data)) == sizeof(data));
close(fd);
}
static int set_mode(struct test_output *o, int fb, int x, int y)
{
int n;
for (n = 0; n < o->count; n++) {
if (fb == 0) {
if (drmModeSetCrtc(drm_fd, o->_crtc[n],
0, 0, 0,
0, 0, 0))
return -1;
} else {
if (drmModeSetCrtc(drm_fd, o->_crtc[n],
fb, x, y,
&o->_connector[n], 1, &o->kmode[n]))
return -1;
}
}
return 0;
}
/* Return mask of completed events. */
static unsigned int run_test_step(struct test_output *o)
{
unsigned int new_fb_id;
/* for funny reasons page_flip returns -EBUSY on disabled crtcs ... */
int expected_einval = o->flags & TEST_MODESET ? -EBUSY : -EINVAL;
unsigned int completed_events = 0;
bool do_flip;
bool do_vblank;
struct vblank_reply vbl_reply;
unsigned int target_seq;
uint32_t handle = 0; /* Suppress GCC warning */
target_seq = o->vblank_state.seq_step;
if (o->flags & TEST_VBLANK_ABSOLUTE)
target_seq += o->vblank_state.last_seq;
/*
* It's possible that we don't have a pending flip here, in case both
* wf-vblank and flip were scheduled and the wf-vblank event was
* delivered earlier. The same applies to vblank events w.r.t flip.
*/
do_flip = (o->flags & TEST_FLIP) && !(o->pending_events & EVENT_FLIP);
do_vblank = (o->flags & TEST_VBLANK) &&
!(o->pending_events & EVENT_VBLANK);
if (o->flags & TEST_DPMS_OFF_OTHERS)
dpms_off_other_outputs(o);
if (o->flags & TEST_WITH_DUMMY_BCS)
emit_dummy_load__bcs(o);
if (o->flags & TEST_WITH_DUMMY_RCS)
emit_dummy_load__rcs(o);
if (!(o->flags & TEST_SINGLE_BUFFER))
o->current_fb_id = !o->current_fb_id;
if (o->flags & TEST_FB_RECREATE)
recreate_fb(o);
new_fb_id = o->fb_ids[o->current_fb_id];
if (o->flags & TEST_FB_BAD_TILING)
new_fb_id = o->fb_ids[2];
if ((o->flags & TEST_VBLANK_EXPIRED_SEQ) &&
!(o->pending_events & EVENT_VBLANK) && o->flip_state.count > 0) {
struct vblank_reply reply;
unsigned int exp_seq;
unsigned long start;
exp_seq = o->flip_state.current_seq;
start = gettime_us();
do_or_die(__wait_for_vblank(TEST_VBLANK_ABSOLUTE |
TEST_VBLANK_BLOCK, o->pipe, exp_seq,
0, &reply));
igt_assert(gettime_us() - start < 500);
igt_assert(reply.sequence == exp_seq);
igt_assert(timercmp(&reply.ts, &o->flip_state.last_ts, ==));
}
if (do_flip && (o->flags & TEST_EINVAL) && o->flip_state.count > 0)
igt_assert(do_page_flip(o, new_fb_id, true) == expected_einval);
if (o->flags & TEST_FB_BAD_TILING)
new_fb_id = o->fb_ids[o->current_fb_id];
if (do_vblank && (o->flags & TEST_EINVAL) && o->vblank_state.count > 0)
igt_assert(do_wait_for_vblank(o, o->pipe, target_seq, &vbl_reply)
== -EINVAL);
if (o->flags & TEST_DPMS_OFF)
set_dpms(o, DRM_MODE_DPMS_OFF);
if (o->flags & TEST_MODESET)
igt_assert(set_mode(o, o->fb_ids[o->current_fb_id], 0, 0) == 0);
if (o->flags & TEST_DPMS)
set_dpms(o, DRM_MODE_DPMS_ON);
printf("."); fflush(stdout);
if (do_flip && (o->flags & TEST_HANG)) {
handle = gem_create(drm_fd, 4096);
hang_gpu(o);
exec_nop(drm_fd, handle);
}
if (do_flip)
do_or_die(do_page_flip(o, new_fb_id, !(o->flags & TEST_NOEVENT)));
if (do_vblank) {
do_or_die(do_wait_for_vblank(o, o->pipe, target_seq,
&vbl_reply));
if (o->flags & TEST_VBLANK_BLOCK) {
event_handler(&o->vblank_state, vbl_reply.sequence,
vbl_reply.ts.tv_sec,
vbl_reply.ts.tv_usec);
completed_events = EVENT_VBLANK;
}
}
if (do_flip && (o->flags & TEST_EBUSY))
igt_assert(do_page_flip(o, new_fb_id, true) == -EBUSY);
if (do_flip && (o->flags & TEST_RMFB))
recreate_fb(o);
/* pan before the flip completes */
if (o->flags & TEST_PAN) {
int count = do_flip ?
o->flip_state.count : o->vblank_state.count;
int x_ofs = count * 10 > o->fb_width - o->kmode[0].hdisplay ? o->fb_width - o->kmode[0].hdisplay : count * 10;
igt_assert_f(set_mode(o, o->fb_ids[o->current_fb_id], x_ofs, 0) == 0,
"failed to pan (%dx%d@%dHz)+%d: %s\n",
o->kmode[0].hdisplay, o->kmode[0].vdisplay, o->kmode[0].vrefresh,
x_ofs, strerror(errno));
}
if (o->flags & TEST_DPMS)
set_dpms(o, DRM_MODE_DPMS_OFF);
if (o->flags & TEST_MODESET && !(o->flags & TEST_RMFB))
igt_assert_f(set_mode(o, 0 /* no fb */, 0, 0),
"failed to disable output: %s\n",
strerror(errno));
if (do_vblank && (o->flags & TEST_EINVAL) && o->vblank_state.count > 0)
igt_assert(do_wait_for_vblank(o, o->pipe, target_seq, &vbl_reply)
== -EINVAL);
if (do_flip && (o->flags & TEST_EINVAL) && !(o->flags & TEST_FB_BAD_TILING))
igt_assert(do_page_flip(o, new_fb_id, true) == expected_einval);
if (do_flip && (o->flags & TEST_HANG)) {
gem_sync(drm_fd, handle);
gem_close(drm_fd, handle);
eat_error_state(o);
}
return completed_events;
}
static void update_state(struct event_state *es)
{
es->last_received_ts = es->current_received_ts;
es->last_ts = es->current_ts;
es->last_seq = es->current_seq;
es->count++;
}
static void update_all_state(struct test_output *o,
unsigned int completed_events)
{
if (completed_events & EVENT_FLIP)
update_state(&o->flip_state);
if (completed_events & EVENT_VBLANK)
update_state(&o->vblank_state);
}
static void connector_find_preferred_mode(uint32_t connector_id, int crtc_idx,
struct test_output *o)
{
struct kmstest_connector_config config;
if (kmstest_get_connector_config(drm_fd, connector_id, 1 << crtc_idx,
&config) < 0) {
o->mode_valid = 0;
return;
}
o->pipe = config.pipe;
o->kconnector[0] = config.connector;
o->kencoder[0] = config.encoder;
o->_crtc[0] = config.crtc->crtc_id;
o->kmode[0] = config.default_mode;
o->mode_valid = 1;
o->fb_width = o->kmode[0].hdisplay;
o->fb_height = o->kmode[0].vdisplay;
}
static bool mode_compatible(const drmModeModeInfo *a, const drmModeModeInfo *b)
{
int d_refresh;
if (a->hdisplay != b->hdisplay)
return false;
if (a->vdisplay != b->vdisplay)
return false;
d_refresh = a->vrefresh - b->vrefresh;
if (d_refresh < -1 || d_refresh > 1)
return false;
return true;
}
static void connector_find_compatible_mode(int crtc_idx0, int crtc_idx1,
struct test_output *o)
{
struct kmstest_connector_config config[2];
drmModeModeInfo *mode[2];
int n, m;
if (kmstest_get_connector_config(drm_fd, o->_connector[0],
1 << crtc_idx0, &config[0]) < 0)
return;
if (kmstest_get_connector_config(drm_fd, o->_connector[1],
1 << crtc_idx1, &config[1]) < 0)
return;
mode[0] = &config[0].default_mode;
mode[1] = &config[1].default_mode;
if (!mode_compatible(mode[0], mode[1])) {
for (n = 0; n < config[0].connector->count_modes; n++) {
mode[0] = &config[0].connector->modes[n];
for (m = 0; m < config[1].connector->count_modes; m++) {
mode[1] = &config[1].connector->modes[m];
if (mode_compatible(mode[0], mode[1]))
goto found;
}
}
/* hope for the best! */
mode[1] = mode[0] = &config[0].default_mode;
}
found:
o->pipe = config[0].pipe;
o->fb_width = mode[0]->hdisplay;
o->fb_height = mode[0]->vdisplay;
o->mode_valid = 1;
o->kconnector[0] = config[0].connector;
o->kencoder[0] = config[0].encoder;
o->_crtc[0] = config[0].crtc->crtc_id;
o->kmode[0] = *mode[0];
o->kconnector[1] = config[1].connector;
o->kencoder[1] = config[1].encoder;
o->_crtc[1] = config[1].crtc->crtc_id;
o->kmode[1] = *mode[1];
}
static void paint_flip_mode(struct kmstest_fb *fb, bool odd_frame)
{
cairo_t *cr = kmstest_get_cairo_ctx(drm_fd, fb);
int width = fb->width;
int height = fb->height;
kmstest_paint_test_pattern(cr, width, height);
if (odd_frame)
cairo_rectangle(cr, width/4, height/2, width/4, height/8);
else
cairo_rectangle(cr, width/2, height/2, width/4, height/8);
cairo_set_source_rgb(cr, 1, 1, 1);
cairo_fill(cr);
igt_assert(!cairo_status(cr));
}
static int
fb_is_bound(struct test_output *o, int fb)
{
int n;
for (n = 0; n < o->count; n++) {
struct drm_mode_crtc mode;
mode.crtc_id = o->_crtc[n];
if (drmIoctl(drm_fd, DRM_IOCTL_MODE_GETCRTC, &mode))
return 0;
if (!mode.mode_valid || mode.fb_id != fb)
return false;
}
return true;
}
static void check_final_state(struct test_output *o, struct event_state *es,
unsigned int elapsed)
{
igt_assert_f(es->count > 0,
"no %s event received\n", es->name);
/* Verify we drop no frames, but only if it's not a TV encoder, since
* those use some funny fake timings behind userspace's back. */
if (o->flags & TEST_CHECK_TS && !analog_tv_connector(o)) {
int expected;
int count = es->count;
count *= es->seq_step;
expected = elapsed * o->kmode[0].vrefresh / (1000 * 1000);
igt_assert_f(count >= expected * 99/100 && count <= expected * 101/100,
"dropped frames, expected %d, counted %d, encoder type %d\n",
expected, count, o->kencoder[0]->encoder_type);
}
}
/*
* Wait until at least one pending event completes. Return mask of completed
* events.
*/
static unsigned int wait_for_events(struct test_output *o)
{
drmEventContext evctx;
struct timeval timeout = { .tv_sec = 3, .tv_usec = 0 };
fd_set fds;
unsigned int event_mask;
int ret;
event_mask = o->pending_events;
igt_assert(event_mask);
memset(&evctx, 0, sizeof evctx);
evctx.version = DRM_EVENT_CONTEXT_VERSION;
evctx.vblank_handler = vblank_handler;
evctx.page_flip_handler = page_flip_handler;
/* make timeout lax with the dummy load */
if (o->flags & (TEST_WITH_DUMMY_BCS | TEST_WITH_DUMMY_RCS))
timeout.tv_sec *= 10;
FD_ZERO(&fds);
FD_SET(drm_fd, &fds);
do {
do {
ret = select(drm_fd + 1, &fds, NULL, NULL, &timeout);
} while (ret < 0 && errno == EINTR);
igt_assert_f(ret > 0,
"select timed out or error (ret %d)\n", ret);
igt_assert_f(!FD_ISSET(0, &fds),
"no fds active, breaking\n");
do_or_die(drmHandleEvent(drm_fd, &evctx));
} while (o->pending_events);
event_mask ^= o->pending_events;
igt_assert(event_mask);
return event_mask;
}
/* Returned the elapsed time in us */
static unsigned event_loop(struct test_output *o, unsigned duration_ms)
{
unsigned long start, end;
start = gettime_us();
while (1) {
unsigned int completed_events;
completed_events = run_test_step(o);
if (o->pending_events)
completed_events |= wait_for_events(o);
check_all_state(o, completed_events);
update_all_state(o, completed_events);
if ((gettime_us() - start) / 1000 >= duration_ms)
break;
}
end = gettime_us();
/* Flush any remaining events */
if (o->pending_events)
wait_for_events(o);
return end - start;
}
static void run_test_on_crtc(struct test_output *o, int crtc_idx, int duration_ms)
{
unsigned elapsed;
connector_find_preferred_mode(o->_connector[0], crtc_idx, o);
if (!o->mode_valid)
return;
last_connector = o->kconnector[0];
fprintf(stdout, "Beginning %s on crtc %d, connector %d\n",
igt_subtest_name(), o->_crtc[0], o->_connector[0]);
if (o->flags & TEST_PAN)
o->fb_width <<= 1;
o->fb_ids[0] = kmstest_create_fb(drm_fd, o->fb_width, o->fb_height,
o->bpp, o->depth, false, &o->fb_info[0]);
o->fb_ids[1] = kmstest_create_fb(drm_fd, o->fb_width, o->fb_height,
o->bpp, o->depth, false, &o->fb_info[1]);
o->fb_ids[2] = kmstest_create_fb(drm_fd, o->fb_width, o->fb_height,
o->bpp, o->depth, true, &o->fb_info[2]);
igt_assert(o->fb_ids[0]);
igt_assert(o->fb_ids[1]);
igt_assert(o->fb_ids[2]);
paint_flip_mode(&o->fb_info[0], false);
paint_flip_mode(&o->fb_info[1], true);
paint_flip_mode(&o->fb_info[2], true);
set_y_tiling(o, 2);
kmstest_dump_mode(&o->kmode[0]);
if (set_mode(o, o->fb_ids[0], 0, 0)) {
/* We may fail to apply the mode if there are hidden
* constraints, such as bandwidth on the third pipe.
*/
igt_assert_f(crtc_idx < 2, "set_mode may only fail on the 3rd pipe\n");
goto out;
}
igt_assert(fb_is_bound(o, o->fb_ids[0]));
/* quiescent the hw a bit so ensure we don't miss a single frame */
if (o->flags & TEST_CHECK_TS)
sleep(1);
igt_assert(do_page_flip(o, o->fb_ids[1], true) == 0);
wait_for_events(o);
o->current_fb_id = 1;
o->flip_state.seq_step = 1;
if (o->flags & TEST_VBLANK_ABSOLUTE)
o->vblank_state.seq_step = 5;
else
o->vblank_state.seq_step = 1;
elapsed = event_loop(o, duration_ms);
if (o->flags & TEST_FLIP && !(o->flags & TEST_NOEVENT))
check_final_state(o, &o->flip_state, elapsed);
if (o->flags & TEST_VBLANK)
check_final_state(o, &o->vblank_state, elapsed);
fprintf(stdout, "\n%s on crtc %d, connector %d: PASSED\n\n",
igt_subtest_name(), o->_crtc[0], o->_connector[0]);
out:
kmstest_remove_fb(drm_fd, &o->fb_info[2]);
kmstest_remove_fb(drm_fd, &o->fb_info[1]);
kmstest_remove_fb(drm_fd, &o->fb_info[0]);
last_connector = NULL;
drmModeFreeEncoder(o->kencoder[0]);
drmModeFreeConnector(o->kconnector[0]);
}
static void run_test_on_crtc_pair(struct test_output *o,
int crtc_idx0, int crtc_idx1, int duration_ms)
{
unsigned elapsed;
int i;
connector_find_compatible_mode(crtc_idx0, crtc_idx1, o);
if (!o->mode_valid)
return;
last_connector = o->kconnector[0];
fprintf(stdout, "Beginning %s on crtc %d:%d, connector %d:%d\n",
igt_subtest_name(), o->_crtc[0], o->_crtc[1], o->_connector[0], o->_connector[1]);
if (o->flags & TEST_PAN)
o->fb_width *= 2;
o->fb_ids[0] = kmstest_create_fb(drm_fd, o->fb_width, o->fb_height,
o->bpp, o->depth, false, &o->fb_info[0]);
o->fb_ids[1] = kmstest_create_fb(drm_fd, o->fb_width, o->fb_height,
o->bpp, o->depth, false, &o->fb_info[1]);
o->fb_ids[2] = kmstest_create_fb(drm_fd, o->fb_width, o->fb_height,
o->bpp, o->depth, true, &o->fb_info[2]);
igt_assert(o->fb_ids[0]);
igt_assert(o->fb_ids[1]);
igt_assert(o->fb_ids[2]);
paint_flip_mode(&o->fb_info[0], false);
paint_flip_mode(&o->fb_info[1], true);
paint_flip_mode(&o->fb_info[2], true);
set_y_tiling(o, 2);
kmstest_dump_mode(&o->kmode[0]);
kmstest_dump_mode(&o->kmode[1]);
if (set_mode(o, o->fb_ids[0], 0, 0)) {
/* We may fail to apply the mode if there are hidden
* constraints, such as bandwidth on the third pipe.
*/
if (0) {
fprintf(stderr, "failed to set mode (%dx%d@%dHz): %s\n",
o->kmode[0].hdisplay, o->kmode[0].vdisplay, o->kmode[0].vrefresh,
strerror(errno));
}
goto out;
}
igt_assert(fb_is_bound(o, o->fb_ids[0]));
/* quiescent the hw a bit so ensure we don't miss a single frame */
if (o->flags & TEST_CHECK_TS)
sleep(1);
igt_assert(do_page_flip(o, o->fb_ids[1], true) == 0);
wait_for_events(o);
o->current_fb_id = 1;
o->flip_state.seq_step = 1;
if (o->flags & TEST_VBLANK_ABSOLUTE)
o->vblank_state.seq_step = 5;
else
o->vblank_state.seq_step = 1;
elapsed = event_loop(o, duration_ms);
if (o->flags & TEST_FLIP && !(o->flags & TEST_NOEVENT))
check_final_state(o, &o->flip_state, elapsed);
if (o->flags & TEST_VBLANK)
check_final_state(o, &o->vblank_state, elapsed);
fprintf(stdout, "\n%s on crtc %d:%d, connector %d:%d: PASSED\n\n",
igt_subtest_name(), o->_crtc[0], o->_crtc[1], o->_connector[0], o->_connector[1]);
out:
kmstest_remove_fb(drm_fd, &o->fb_info[2]);
kmstest_remove_fb(drm_fd, &o->fb_info[1]);
kmstest_remove_fb(drm_fd, &o->fb_info[0]);
last_connector = NULL;
for (i = 0; i < o->count; i++) {
drmModeFreeEncoder(o->kencoder[i]);
drmModeFreeConnector(o->kconnector[i]);
}
}
static int run_test(int duration, int flags)
{
struct test_output o;
int i, n, modes = 0;
resources = drmModeGetResources(drm_fd);
igt_assert(resources);
/* Count output configurations to scale test runtime. */
for (i = 0; i < resources->count_connectors; i++) {
for (n = 0; n < resources->count_crtcs; n++) {
memset(&o, 0, sizeof(o));
o.count = 1;
o._connector[0] = resources->connectors[i];
o.flags = flags;
o.flip_state.name = "flip";
o.vblank_state.name = "vblank";
o.bpp = 32;
o.depth = 24;
connector_find_preferred_mode(o._connector[0], n, &o);
if (o.mode_valid)
modes++;
}
}
igt_assert(modes);
duration = duration * 1000 / modes;
duration = duration < 500 ? 500 : duration;
/* Find any connected displays */
for (i = 0; i < resources->count_connectors; i++) {
for (n = 0; n < resources->count_crtcs; n++) {
memset(&o, 0, sizeof(o));
o.count = 1;
o._connector[0] = resources->connectors[i];
o.flags = flags;
o.flip_state.name = "flip";
o.vblank_state.name = "vblank";
o.bpp = 32;
o.depth = 24;
run_test_on_crtc(&o, n, duration);
}
}
drmModeFreeResources(resources);
return 1;
}
static int run_pair(int duration, int flags)
{
struct test_output o;
int i, j, m, n, modes = 0;
resources = drmModeGetResources(drm_fd);
igt_assert(resources);
/* Find a pair of connected displays */
for (i = 0; i < resources->count_connectors; i++) {
for (n = 0; n < resources->count_crtcs; n++) {
for (j = i + 1; j < resources->count_connectors; j++) {
for (m = n + 1; m < resources->count_crtcs; m++) {
memset(&o, 0, sizeof(o));
o.count = 2;
o._connector[0] = resources->connectors[i];
o._connector[1] = resources->connectors[j];
o.flags = flags;
o.flip_state.name = "flip";
o.vblank_state.name = "vblank";
o.bpp = 32;
o.depth = 24;
connector_find_compatible_mode(n, m, &o);
if (o.mode_valid)
modes++;
}
}
}
}
/* If we have fewer than 2 connected outputs then we won't have any
* configuration at all. So skip in that case. */
igt_require(modes);
duration = duration * 1000 / modes;
duration = duration < 500 ? 500 : duration;
/* Find a pair of connected displays */
for (i = 0; i < resources->count_connectors; i++) {
for (n = 0; n < resources->count_crtcs; n++) {
for (j = i + 1; j < resources->count_connectors; j++) {
for (m = n + 1; m < resources->count_crtcs; m++) {
memset(&o, 0, sizeof(o));
o.count = 2;
o._connector[0] = resources->connectors[i];
o._connector[1] = resources->connectors[j];
o.flags = flags;
o.flip_state.name = "flip";
o.vblank_state.name = "vblank";
o.bpp = 32;
o.depth = 24;
run_test_on_crtc_pair(&o, n, m, duration);
}
}
}
}
drmModeFreeResources(resources);
return 1;
}
static void get_timestamp_format(void)
{
uint64_t cap_mono;
int ret;
ret = drmGetCap(drm_fd, DRM_CAP_TIMESTAMP_MONOTONIC, &cap_mono);
igt_assert(ret == 0 || errno == EINVAL);
monotonic_timestamp = ret == 0 && cap_mono == 1;
printf("Using %s timestamps\n",
monotonic_timestamp ? "monotonic" : "real");
}
static void kms_flip_exit_handler(int sig)
{
if (last_connector)
set_connector_dpms(last_connector, DRM_MODE_DPMS_ON);
}
int main(int argc, char **argv)
{
struct {
int duration;
int flags;
const char *name;
} tests[] = {
{ 30, TEST_VBLANK, "wf_vblank" },
{ 30, TEST_VBLANK | TEST_CHECK_TS, "wf_vblank-ts-check" },
{ 30, TEST_VBLANK | TEST_VBLANK_BLOCK | TEST_CHECK_TS,
"blocking-wf_vblank" },
{ 30, TEST_VBLANK | TEST_VBLANK_ABSOLUTE,
"absolute-wf_vblank" },
{ 30, TEST_VBLANK | TEST_VBLANK_BLOCK | TEST_VBLANK_ABSOLUTE,
"blocking-absolute-wf_vblank" },
{ 60, TEST_VBLANK | TEST_DPMS | TEST_EINVAL, "wf_vblank-vs-dpms" },
{ 60, TEST_VBLANK | TEST_DPMS | TEST_WITH_DUMMY_BCS,
"bcs-wf_vblank-vs-dpms" },
{ 60, TEST_VBLANK | TEST_DPMS | TEST_WITH_DUMMY_RCS,
"rcs-wf_vblank-vs-dpms" },
{ 60, TEST_VBLANK | TEST_MODESET | TEST_EINVAL, "wf_vblank-vs-modeset" },
{ 60, TEST_VBLANK | TEST_MODESET | TEST_WITH_DUMMY_BCS,
"bcs-wf_vblank-vs-modeset" },
{ 60, TEST_VBLANK | TEST_MODESET | TEST_WITH_DUMMY_RCS,
"rcs-wf_vblank-vs-modeset" },
{ 30, TEST_FLIP | TEST_EBUSY , "plain-flip" },
{ 30, TEST_FLIP | TEST_CHECK_TS | TEST_EBUSY , "plain-flip-ts-check" },
{ 30, TEST_FLIP | TEST_CHECK_TS | TEST_EBUSY | TEST_FB_RECREATE,
"plain-flip-fb-recreate" },
{ 30, TEST_FLIP | TEST_EBUSY | TEST_RMFB | TEST_MODESET , "flip-vs-rmfb" },
{ 60, TEST_FLIP | TEST_DPMS | TEST_EINVAL, "flip-vs-dpms" },
{ 60, TEST_FLIP | TEST_DPMS | TEST_WITH_DUMMY_BCS, "bcs-flip-vs-dpms" },
{ 60, TEST_FLIP | TEST_DPMS | TEST_WITH_DUMMY_RCS, "rcs-flip-vs-dpms" },
{ 30, TEST_FLIP | TEST_PAN, "flip-vs-panning" },
{ 60, TEST_FLIP | TEST_PAN | TEST_WITH_DUMMY_BCS, "bcs-flip-vs-panning" },
{ 60, TEST_FLIP | TEST_PAN | TEST_WITH_DUMMY_RCS, "rcs-flip-vs-panning" },
{ 60, TEST_FLIP | TEST_MODESET | TEST_EINVAL, "flip-vs-modeset" },
{ 60, TEST_FLIP | TEST_MODESET | TEST_WITH_DUMMY_BCS, "bcs-flip-vs-modeset" },
{ 60, TEST_FLIP | TEST_MODESET | TEST_WITH_DUMMY_RCS, "rcs-flip-vs-modeset" },
{ 30, TEST_FLIP | TEST_VBLANK_EXPIRED_SEQ,
"flip-vs-expired-vblank" },
{ 30, TEST_FLIP | TEST_VBLANK | TEST_VBLANK_ABSOLUTE |
TEST_CHECK_TS, "flip-vs-absolute-wf_vblank" },
{ 30, TEST_FLIP | TEST_VBLANK | TEST_CHECK_TS,
"flip-vs-wf_vblank" },
{ 30, TEST_FLIP | TEST_VBLANK | TEST_VBLANK_BLOCK |
TEST_CHECK_TS, "flip-vs-blocking-wf-vblank" },
{ 30, TEST_FLIP | TEST_MODESET | TEST_HANG | TEST_NOEVENT, "flip-vs-modeset-vs-hang" },
{ 30, TEST_FLIP | TEST_PAN | TEST_HANG, "flip-vs-panning-vs-hang" },
{ 1, TEST_FLIP | TEST_EINVAL | TEST_FB_BAD_TILING, "flip-vs-bad-tiling" },
{ 1, TEST_DPMS_OFF | TEST_MODESET | TEST_FLIP,
"flip-vs-dpms-off-vs-modeset" },
{ 1, TEST_DPMS_OFF | TEST_MODESET | TEST_FLIP | TEST_SINGLE_BUFFER,
"single-buffer-flip-vs-dpms-off-vs-modeset" },
{ 30, TEST_FLIP | TEST_NO_2X_OUTPUT | TEST_DPMS_OFF_OTHERS , "dpms-off-confusion" },
};
int i;
igt_subtest_init(argc, argv);
igt_skip_on_simulation();
igt_fixture {
drm_fd = drm_open_any();
do_or_die(igt_set_vt_graphics_mode());
do_or_die(igt_install_exit_handler(kms_flip_exit_handler));
get_timestamp_format();
bufmgr = drm_intel_bufmgr_gem_init(drm_fd, 4096);
devid = intel_get_drm_devid(drm_fd);
batch = intel_batchbuffer_alloc(bufmgr, devid);
}
for (i = 0; i < sizeof(tests) / sizeof (tests[0]); i++) {
igt_subtest(tests[i].name)
run_test(tests[i].duration, tests[i].flags);
if (tests[i].flags & TEST_NO_2X_OUTPUT)
continue;
igt_subtest_f( "2x-%s", tests[i].name)
run_pair(tests[i].duration, tests[i].flags);
}
igt_fork_signal_helper();
for (i = 0; i < sizeof(tests) / sizeof (tests[0]); i++) {
/* relative blocking vblank waits that get constantly interrupt
* take forver. So don't do them. */
if ((tests[i].flags & TEST_VBLANK_BLOCK) &&
!(tests[i].flags & TEST_VBLANK_ABSOLUTE))
continue;
igt_subtest_f( "%s-interruptible", tests[i].name)
run_test(tests[i].duration, tests[i].flags);
if (tests[i].flags & TEST_NO_2X_OUTPUT)
continue;
igt_subtest_f( "2x-%s-interruptible", tests[i].name)
run_pair(tests[i].duration, tests[i].flags);
}
igt_stop_signal_helper();
igt_fixture
close(drm_fd);
igt_exit();
}