From 46ff17559b1369f0fe7dca000f51f077ffd1dae5 Mon Sep 17 00:00:00 2001 From: Rob Clark Date: Mon, 23 May 2016 15:53:23 -0400 Subject: [PATCH] freedreno/ir3: disable cp for indirect src's The variable-indexing tests always had a few random fails, which I usually couldn't reproduce when running tests manually. Somehow recently this got a lot worse. I ported a couple of the shaders to GLES to see what blob does, and it also seems to be avoiding to cp indirect srcs. So I guess indirect w/ instructions other than cat1 (mov) are not totally reliable. Let's just switch that off until this is better understood. Signed-off-by: Rob Clark --- src/gallium/drivers/freedreno/ir3/ir3_cp.c | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/src/gallium/drivers/freedreno/ir3/ir3_cp.c b/src/gallium/drivers/freedreno/ir3/ir3_cp.c index 6e71bee..57c37e2 100644 --- a/src/gallium/drivers/freedreno/ir3/ir3_cp.c +++ b/src/gallium/drivers/freedreno/ir3/ir3_cp.c @@ -102,6 +102,15 @@ static bool valid_flags(struct ir3_instruction *instr, unsigned n, (flags & IR3_REG_RELATIV)) return false; + /* TODO it seems to *mostly* work to cp RELATIV, except we get some + * intermittent piglit variable-indexing fails. Newer blob driver + * doesn't seem to cp these. Possibly this is hw workaround? Not + * sure, but until that is understood better, lets just switch off + * cp for indirect src's: + */ + if (flags & IR3_REG_RELATIV) + return false; + /* clear flags that are 'ok' */ switch (opc_cat(instr->opc)) { case 1: -- 2.7.4