> -----Original Message-----
> From: Thomas Monjalon <[email protected]>
> Sent: Monday, June 12, 2023 10:15 PM
> To: Zhang, Qi Z <[email protected]>
> Cc: Zeng, ZhichaoX <[email protected]>; [email protected]; Richardson,
> Bruce <[email protected]>; Konstantin Ananyev
> <[email protected]>; Wu, Jingjing <[email protected]>;
> Xing, Beilei <[email protected]>; Wu, Wenjun1 <[email protected]>;
> Gao, DaxueX <[email protected]>; [email protected];
> Mcnamara, John <[email protected]>; Zhang, Helin
> <[email protected]>
> Subject: Re: [PATCH v2] net/iavf: fix compilation warnings
>
> 12/06/2023 14:29, Zhang, Qi Z:
> > From: Thomas Monjalon <[email protected]>
> > > 12/06/2023 10:32, Gao, DaxueX:
> > > > From: Zeng, ZhichaoX <[email protected]>
> > > > > Use fallthrough comments to ignore compilation warnings instead
> > > > > of diagnostic pragmas, which are not recognized by old GCC(below
> GCC7).
> > > > >
> > > > > Fixes: 95a1b0b23140 ("net/iavf: support Rx timestamp offload on
> > > > > SSE")
> > > > > Fixes: 4f58266da4d3 ("net/iavf: support Rx timestamp offload on
> > > > > AVX2")
> > > > > Fixes: d65eadb9e178 ("net/iavf: support Rx timestamp offload on
> > > > > AVX512")
> > > > > Signed-off-by: Zhichao Zeng <[email protected]>
> > > > Tested-by: Daxue Gao <[email protected]>
> > >
> > > Please squash the fix in the original commits above, so we won't
> > > have any compilation issue when doing "git bisect".
> > >
> > > In the meantime, I am pulling the branch next-net-intel without the
> > > Rx timestamp support for SSE/AVX2/AVX512.
> > >
> >
> > Squashed to original patches in dpdk-next-net-intel.
>
> It is still not compiling:
>
> In function '_mm_testz_si128',
> inlined from '_recv_raw_pkts_vec_flex_rxd'
> at ../../dpdk/drivers/net/iavf/iavf_rxtx_vec_sse.c:1070:12:
> /usr/lib/gcc/x86_64-pc-linux-gnu/13.1.1/include/smmintrin.h:69:10: error:
> 'mask_low' may be used uninitialized [-Werror=maybe-uninitialized]
> 69 | return __builtin_ia32_ptestz128 ((__v2di)__M, (__v2di)__V);
> | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> ../../dpdk/drivers/net/iavf/iavf_rxtx_vec_sse.c: In function
> '_recv_raw_pkts_vec_flex_rxd':
> ../../dpdk/drivers/net/iavf/iavf_rxtx_vec_sse.c:1058:49: note: 'mask_low' was
> declared here
> 1058 | __m128i mask_low;
> | ^~~~~~~~
> In function '_mm_testz_si128',
> inlined from '_recv_raw_pkts_vec_flex_rxd'
> at ../../dpdk/drivers/net/iavf/iavf_rxtx_vec_sse.c:1072:13:
> /usr/lib/gcc/x86_64-pc-linux-gnu/13.1.1/include/smmintrin.h:69:10: error:
> 'mask_high' may be used uninitialized [-Werror=maybe-uninitialized]
> 69 | return __builtin_ia32_ptestz128 ((__v2di)__M, (__v2di)__V);
> | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> ../../dpdk/drivers/net/iavf/iavf_rxtx_vec_sse.c: In function
> '_recv_raw_pkts_vec_flex_rxd':
> ../../dpdk/drivers/net/iavf/iavf_rxtx_vec_sse.c:1059:49: note: 'mask_high' was
> declared here
> 1059 | __m128i mask_high;
> | ^~~~~~~~~
> cc1: all warnings being treated as errors
>
>
The question is if "apply patch failure" happens, the CI build will not be
executed and we may not able to figure out the error early, any BKM we should
follow? Thanks.
