From: José Fonseca Date: Fri, 24 Jan 2014 14:25:46 +0000 (+0000) Subject: gallivm: Workaround http://llvm.org/PR18600 X-Git-Url: https://git.libre-soc.org/?a=commitdiff_plain;h=2eddf91faf4abf0b3d7b81a0a486fd46d5acd1ea;p=mesa.git gallivm: Workaround http://llvm.org/PR18600 We have code generation paths that carry out swizzles of AoS vectors via bitwise shifts, as these tend to generate more efficient code than straightforward byte shuffles. But when the input is a constant the additional bitwise arithmetic operations somehow don't really get constant propagated properly, evenutally causing assertion failure in InstCombine pass. Therefore avoid the bug by using the trivial shuffles for constant inputs. Although the sample LLVM IR can cause a crash with any LLVM version, this was only seen in practice with LLVM 3.2. Reviewed-by: Matthew McClure Reviewed-by: Roland Scheidegger --- diff --git a/src/gallium/auxiliary/gallivm/lp_bld_swizzle.c b/src/gallium/auxiliary/gallivm/lp_bld_swizzle.c index 79116bcf391..9557e1c10de 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_swizzle.c +++ b/src/gallium/auxiliary/gallivm/lp_bld_swizzle.c @@ -180,7 +180,8 @@ lp_build_swizzle_scalar_aos(struct lp_build_context *bld, /* XXX: SSE3 has PSHUFB which should be better than bitmasks, but forcing * using shuffles here actually causes worst results. More investigation is * needed. */ - if (type.width >= 16) { + if (LLVMIsConstant(a) || + type.width >= 16) { /* * Shuffle. */ @@ -398,7 +399,8 @@ lp_build_swizzle_aos(struct lp_build_context *bld, } } - if (type.width >= 16) { + if (LLVMIsConstant(a) || + type.width >= 16) { /* * Shuffle. */