From: Tiago Vignatti <tiago.vignatti@intel.com>
To: dri-devel@lists.freedesktop.org
Cc: daniel.vetter@ffwll.ch, intel-gfx@lists.freedesktop.org,
daniel.thompson@linaro.org
Subject: [PATCH i-g-t 5/5] tests/kms_mmap_write_crc: Demonstrate the need for end_cpu_access
Date: Tue, 11 Aug 2015 17:59:28 -0300 [thread overview]
Message-ID: <1439326768-5611-10-git-send-email-tiago.vignatti@intel.com> (raw)
In-Reply-To: <1439326768-5611-1-git-send-email-tiago.vignatti@intel.com>
It requires i915 changes to add end_cpu_access().
Signed-off-by: Tiago Vignatti <tiago.vignatti@intel.com>
---
tests/kms_mmap_write_crc.c | 63 ++++++++++++++++++++++++++++++++++++++++------
1 file changed, 55 insertions(+), 8 deletions(-)
diff --git a/tests/kms_mmap_write_crc.c b/tests/kms_mmap_write_crc.c
index e24d535..59ac9e7 100644
--- a/tests/kms_mmap_write_crc.c
+++ b/tests/kms_mmap_write_crc.c
@@ -67,6 +67,24 @@ static char *dmabuf_mmap_framebuffer(int drm_fd, struct igt_fb *fb)
return ptr;
}
+static void dmabuf_sync_start(void)
+{
+ struct dma_buf_sync sync_start;
+
+ memset(&sync_start, 0, sizeof(sync_start));
+ sync_start.flags = DMA_BUF_SYNC_START | DMA_BUF_SYNC_RW;
+ do_ioctl(dma_buf_fd, DMA_BUF_IOCTL_SYNC, &sync_start);
+}
+
+static void dmabuf_sync_end(void)
+{
+ struct dma_buf_sync sync_end;
+
+ memset(&sync_end, 0, sizeof(sync_end));
+ sync_end.flags = DMA_BUF_SYNC_END | DMA_BUF_SYNC_RW;
+ do_ioctl(dma_buf_fd, DMA_BUF_IOCTL_SYNC, &sync_end);
+}
+
static void test_begin_access(data_t *data)
{
igt_display_t *display = &data->display;
@@ -103,14 +121,11 @@ static void test_begin_access(data_t *data)
caching = gem_get_caching(data->drm_fd, fb->gem_handle);
igt_assert(caching == I915_CACHING_NONE || caching == I915_CACHING_DISPLAY);
- // Uncomment the following for flush and the crc check next passes. It
- // requires the kernel counter-part of it implemented obviously.
- // {
- // struct dma_buf_sync sync_start;
- // memset(&sync_start, 0, sizeof(sync_start));
- // sync_start.flags = DMA_BUF_SYNC_START | DMA_BUF_SYNC_RW;
- // do_ioctl(dma_buf_fd, DMA_BUF_IOCTL_SYNC, &sync_start);
- // }
+ /*
+ * firstly demonstrate the need for DMA_BUF_SYNC_START ("begin_cpu_access")
+ */
+
+ dmabuf_sync_start();
/* use dmabuf pointer to make the other fb all white too */
buf = malloc(fb->size);
@@ -126,6 +141,38 @@ static void test_begin_access(data_t *data)
/* check that the crc is as expected, which requires that caches got flushed */
igt_pipe_crc_collect_crc(data->pipe_crc, &crc);
igt_assert_crc_equal(&crc, &data->ref_crc);
+
+ /*
+ * now demonstrates the need for DMA_BUF_SYNC_END ("end_cpu_access")
+ */
+
+ /* start over, writing non-white to the fb again and flip to it to make it
+ * fully flushed */
+ cr = igt_get_cairo_ctx(data->drm_fd, fb);
+ igt_paint_test_pattern(cr, fb->width, fb->height);
+ cairo_destroy(cr);
+
+ igt_plane_set_fb(data->primary, fb);
+ igt_display_commit(display);
+
+ /* sync start, to move to CPU domain */
+ dmabuf_sync_start();
+
+ /* use dmabuf pointer in the same fb to make it all white */
+ buf = malloc(fb->size);
+ igt_assert(buf != NULL);
+ memset(buf, 0xff, fb->size);
+ memcpy(ptr, buf, fb->size);
+ free(buf);
+
+ /* there's an implicit flush in set_fb() as well (to set to the GTT domain),
+ * so if we don't do it and instead write directly into the fb as it is the
+ * scanout, that should demonstrate the need for end_cpu_access */
+ dmabuf_sync_end();
+
+ /* check that the crc is as expected, which requires that caches got flushed */
+ igt_pipe_crc_collect_crc(data->pipe_crc, &crc);
+ igt_assert_crc_equal(&crc, &data->ref_crc);
}
static bool prepare_crtc(data_t *data)
--
2.1.0
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/dri-devel
prev parent reply other threads:[~2015-08-11 20:59 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-08-11 20:59 [PATCH v3] mmap on the dma-buf directly Tiago Vignatti
2015-08-11 20:59 ` [PATCH 1/4] drm: prime: Honour O_RDWR during prime-handle-to-fd Tiago Vignatti
2015-08-11 20:59 ` [PATCH 2/4] dma-buf: Add ioctls to allow userspace to flush Tiago Vignatti
2015-08-12 14:20 ` Sumit Semwal
2015-08-11 20:59 ` [PATCH 3/4] drm/i915: Implement end_cpu_access Tiago Vignatti
2015-08-11 20:59 ` [PATCH 4/4] drm/i915: Use CPU mapping for userspace dma-buf mmap() Tiago Vignatti
2015-08-11 22:20 ` Chris Wilson
2015-08-12 13:28 ` Daniel Vetter
2015-08-11 20:59 ` [PATCH i-g-t 1/5] prime_mmap: Add new test for calling mmap() on dma-buf fds Tiago Vignatti
2015-08-11 20:59 ` [PATCH i-g-t 2/5] prime_mmap: Fix a few misc stuff Tiago Vignatti
2015-08-11 20:59 ` [PATCH i-g-t 3/5] prime_mmap: Add basic tests to write in a bo using CPU Tiago Vignatti
2015-08-11 20:59 ` [PATCH i-g-t 4/5] tests: Add kms_mmap_write_crc for cache coherency tests Tiago Vignatti
2015-08-11 20:59 ` Tiago Vignatti [this message]
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=1439326768-5611-10-git-send-email-tiago.vignatti@intel.com \
--to=tiago.vignatti@intel.com \
--cc=daniel.thompson@linaro.org \
--cc=daniel.vetter@ffwll.ch \
--cc=dri-devel@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