From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mgw.gov.kz (mgw.gov.kz [89.218.88.242]) by dpdk.org (Postfix) with ESMTP id 9E557592C for ; Tue, 18 Nov 2014 15:56:58 +0100 (CET) Received: from mgw.gov.kz (mx.ctsat.kz [178.89.4.95]) by mgw.gov.kz with ESMTP id sAIF7HnZ018314-sAIF7Hnb018314 (version=TLSv1.0 cipher=AES128-SHA bits=128 verify=NO); Tue, 18 Nov 2014 21:07:18 +0600 Received: from EXCASHUB1.rgp.local (192.168.40.51) by EdgeForefront.rgp.local (192.168.40.59) with Microsoft SMTP Server (TLS) id 14.2.247.3; Tue, 18 Nov 2014 21:07:06 +0600 Received: from [192.168.35.15] (192.168.35.15) by excashub1.rgp.local (192.168.40.48) with Microsoft SMTP Server (TLS) id 14.2.247.3; Tue, 18 Nov 2014 21:07:17 +0600 Message-ID: <546B607B.9030808@sts.kz> Date: Tue, 18 Nov 2014 21:06:35 +0600 From: Yerden Zhumabekov User-Agent: Mozilla/5.0 (Windows NT 6.3; WOW64; rv:31.0) Gecko/20100101 Thunderbird/31.2.0 MIME-Version: 1.0 To: Neil Horman , "dev@dpdk.org" References: <1409724351-23786-1-git-send-email-e_zhumabekov@sts.kz> <20141118144138.GB32375@hmsreliant.think-freely.org> In-Reply-To: <20141118144138.GB32375@hmsreliant.think-freely.org> Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable X-Originating-IP: [192.168.35.15] X-FEAS-SYSTEM-WL: e_zhumabekov@sts.kz Subject: Re: [dpdk-dev] [PATCH v4 3/5] hash: add fallback to software CRC32 implementation X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: patches and discussions about DPDK List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 18 Nov 2014 14:56:59 -0000 18.11.2014 20:41, Neil Horman =D0=BF=D0=B8=D1=88=D0=B5=D1=82: > On Tue, Nov 18, 2014 at 08:03:40PM +0600, Yerden Zhumabekov wrote: >> Initially, SSE4.2 support is detected via CPUID instruction. >> >> Added rte_hash_crc_set_alg() function to detect and set CRC32 >> implementation if necessary. SSE4.2 is allowed by default. If it's >> not available, fall back to sw implementation. >> >> Best available algorithm is detected upon application startup >> through the constructor function rte_hash_crc_try_sse442(). >> >> Signed-off-by: Yerden Zhumabekov >> --- >> lib/librte_hash/rte_hash_crc.h | 53 +++++++++++++++++++++++++++++++= +++++++-- >> 1 file changed, 51 insertions(+), 2 deletions(-) >> >> diff --git a/lib/librte_hash/rte_hash_crc.h b/lib/librte_hash/rte_hash= _crc.h >> index 15f687a..332ed99 100644 >> --- a/lib/librte_hash/rte_hash_crc.h >> +++ b/lib/librte_hash/rte_hash_crc.h >> @@ -45,7 +45,11 @@ extern "C" { >> #endif >> =20 >> #include >> +#ifdef RTE_MACHINE_CPUFLAG_SSE4_2 >> #include >> +#endif >> +#include >> +#include >> =20 >> /* Lookup tables for software implementation of CRC32C */ >> static uint32_t crc32c_tables[8][256] =3D {{ >> @@ -363,8 +367,41 @@ crc32c_2words(uint64_t data, uint32_t init_val) >> return crc; >> } >> =20 >> +enum crc32_alg_t { >> + CRC32_SW =3D 0, >> + CRC32_SSE42 >> +}; >> + >> +static enum crc32_alg_t crc32_alg; >> + >> +/** >> + * Allow or disallow use of SSE4.2 instrinsics for CRC32 hash >> + * calculation. >> + * >> + * @param flag >> + * unsigned integer flag >> + * - (CRC32_SW) Don't use SSE4.2 intrinsics >> + * - (CRC32_SSE42) Use SSE4.2 intrinsics if available, set by defau= lt >> + */ >> +static inline void >> +rte_hash_crc_set_alg(enum crc32_alg_t alg) >> +{ >> + int sse42_supp =3D rte_cpu_get_flag_enabled(RTE_CPUFLAG_SSE4_2); >> + enum crc32_alg_t alg_supp =3D sse42_supp ? CRC32_SSE42 : CRC32_SW; >> + crc32_alg =3D (alg =3D=3D CRC32_SSE42) ? alg_supp : CRC32_SW; >> +} >> + >> +/* Best available algorithm is detected via CPUID instruction */ >> +static inline void __attribute__((constructor)) >> +rte_hash_crc_try_sse42(void) >> +{ >> + rte_hash_crc_set_alg(CRC32_SSE42); >> +} >> + >> /** >> * Use single crc32 instruction to perform a hash on a 4 byte value. >> + * Fall back to software crc32 implementation in case SSE4.2 is >> + * not supported >> * >> * @param data >> * Data to perform hash on. >> @@ -376,11 +413,18 @@ crc32c_2words(uint64_t data, uint32_t init_val) >> static inline uint32_t >> rte_hash_crc_4byte(uint32_t data, uint32_t init_val) >> { >> - return _mm_crc32_u32(init_val, data); >> +#ifdef RTE_MACHINE_CPUFLAG_SSE4_2 >> + if (likely(crc32_alg =3D=3D CRC32_SSE42)) >> + return _mm_crc32_u32(init_val, data); >> +#endif > you don't really need these ifdefs here anymore given that you have a > constructor to do the algorithm selection. In fact you need to remove = them, in > the event you build on a system that doesn't support SSE42, but run on = a system > that does. Originally, I thought so as well. I wrote the code without these ifdefs, but it didn't compile on my machine which doesn't support SSE4.2. Error was triggered by nmmintrin.h which has a check for respective GCC extension. So I think these ifdefs are indeed required. --=20 Sincerely, Yerden Zhumabekov State Technical Service Astana, KZ