Re: [PATCH 5/6] OMAPDSS: DISPC: move fifo threhold calc to dispc.c

On Sat, 2012-01-14 at 01:30 +0530, Archit wrote:

Hi,

On Friday 13 January 2012 05:16 PM, Tomi Valkeinen wrote:

quoted

Move fifo threshold calculation into dispc.c, as the thresholds are
really dispc internal thing.

Signed-off-by: Tomi Valkeinen<redacted>

<snip>

quoted

diff --git a/drivers/video/omap2/dss/dsi.c b/drivers/video/omap2/dss/dsi.c
index 511ae2a..1cbb7a5 100644
--- a/drivers/video/omap2/dss/dsi.c
+++ b/drivers/video/omap2/dss/dsi.c

@@ -4524,14 +4524,6 @@ int omapdss_dsi_enable_te(struct omap_dss_device *dssdev, bool enable)
  }
  EXPORT_SYMBOL(omapdss_dsi_enable_te);

-void dsi_get_overlay_fifo_thresholds(enum omap_plane plane,
-		u32 fifo_size, u32 burst_size,
-		u32 *fifo_low, u32 *fifo_high)
-{
-	*fifo_high = fifo_size - burst_size;
-	*fifo_low = fifo_size - burst_size * 2;
-}

We are removing the special treatment for overlays connected to DSI done 
before. Won't this cause the issues you saw with DSI in OMAP3?

That's true. I had it in mind at some point, but I seem to have
forgotten it.

The problem with OMAP3, DSI and fifo thresholds was never cleared, and I
haven't seen an errata about it, so there's a slim chance that it was
only a problem with the particular setup.

Are you back at the office yet? If I recall right, you had an OMAP3 DSI
cmd mode board?

Anyway, I guess it's safest if I add a hack there, which tunes the
thresholds a bit differently for OMAP3 DSI.

 Tomi