Hi Zoltan, > -----Original Message----- > From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Zoltan Kiss > Sent: Saturday, September 26, 2015 6:59 AM > To: dev at dpdk.org > Subject: [dpdk-dev] [PATCH] acl: fix target arch detection > > This test selects AVX2 code even if the target architecture doesn't support > it. > > Signed-off-by: Zoltan Kiss <zoltan.kiss at linaro.org> > --- > diff --git a/lib/librte_acl/Makefile b/lib/librte_acl/Makefile > index 46acc2b..17a9f96 100644 > --- a/lib/librte_acl/Makefile > +++ b/lib/librte_acl/Makefile > @@ -57,7 +57,7 @@ CFLAGS_acl_run_sse.o += -msse4.1 > # then add support for AVX2 classify method. > # > > -CC_AVX2_SUPPORT=$(shell $(CC) -march=core-avx2 -dM -E - </dev/null 2>&1 | \ > +CC_AVX2_SUPPORT=$(shell $(CC) $(MACHINE_FLAGS) -dM -E - </dev/null 2>&1 | \ > grep -q AVX2 && echo 1) > > ifeq ($(CC_AVX2_SUPPORT), 1)
The purpose of that code is to check does compiler supports AVX2 or not. If it does, even if selected target doesn't support AVX2, we still compile in AVX2 version of acl_classify (same for SSE4). Then at startup, rte_acl_init() selects the highest supported on that box implementation (scalar/sse/avx2). So NACK. Konstantin BTW, why MACHINE_FLAGS? Shouldn't it be MACHINE_CFLAGS, instead?
