Tegra20 has a high-priority-request control that allows to configure
when display's memory client should perform read requests with a higher
priority (Tegra30+ uses other means like Latency Allowance).

This patch changes the controls configuration in order to get a more
aggressive memory prefetching, which allows to reliably avoid FIFO
underflow when running on a lower memory frequency. This allow us
safely drop the memory bandwidth requirement by about two times in a
most popular use-cases (only one display active, video overlay inactive,
no scaling is done).

Signed-off-by: Dmitry Osipenko <dig...@gmail.com>
---
 drivers/gpu/drm/tegra/dc.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/gpu/drm/tegra/dc.c b/drivers/gpu/drm/tegra/dc.c
index b540ac6ffdc4..564af302a965 100644
--- a/drivers/gpu/drm/tegra/dc.c
+++ b/drivers/gpu/drm/tegra/dc.c
@@ -1980,12 +1980,12 @@ static void tegra_crtc_atomic_enable(struct drm_crtc 
*crtc,
                tegra_dc_writel(dc, value, DC_CMD_INT_POLARITY);
 
                /* initialize timer */
-               value = CURSOR_THRESHOLD(0) | WINDOW_A_THRESHOLD(0x20) |
-                       WINDOW_B_THRESHOLD(0x20) | WINDOW_C_THRESHOLD(0x20);
+               value = CURSOR_THRESHOLD(0) | WINDOW_A_THRESHOLD(0x70) |
+                       WINDOW_B_THRESHOLD(0x30) | WINDOW_C_THRESHOLD(0x70);
                tegra_dc_writel(dc, value, DC_DISP_DISP_MEM_HIGH_PRIORITY);
 
-               value = CURSOR_THRESHOLD(0) | WINDOW_A_THRESHOLD(1) |
-                       WINDOW_B_THRESHOLD(1) | WINDOW_C_THRESHOLD(1);
+               value = CURSOR_THRESHOLD(0) | WINDOW_A_THRESHOLD(0) |
+                       WINDOW_B_THRESHOLD(0) | WINDOW_C_THRESHOLD(0);
                tegra_dc_writel(dc, value, 
DC_DISP_DISP_MEM_HIGH_PRIORITY_TIMER);
 
                value = VBLANK_INT | WIN_A_UF_INT | WIN_B_UF_INT | WIN_C_UF_INT 
|
-- 
2.25.1

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

Reply via email to