From mboxrd@z Thu Jan 1 00:00:00 1970 From: Neil Horman Subject: Re: [PATCH 0/2] rewritten rte_hash_crc() call Date: Fri, 14 Nov 2014 13:41:57 -0500 Message-ID: <20141114184157.GF19147@hmsreliant.think-freely.org> References: <1409724351-23786-1-git-send-email-e_zhumabekov@sts.kz> <2916837.QJI9btJm0N@xps13> <20141114005211.GC14230@localhost.localdomain> <5465AC00.1070602@sts.kz> <20141114113327.GA19147@hmsreliant.think-freely.org> <5465EE3F.2010404@sts.kz> <20141114135308.GD19147@hmsreliant.think-freely.org> <5466313B.6040407@sts.kz> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Cc: dev-VfR2kkLFssw@public.gmane.org To: Yerden Zhumabekov Return-path: Content-Disposition: inline In-Reply-To: <5466313B.6040407-8EHiFRVJVgQ@public.gmane.org> List-Id: patches and discussions about DPDK List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces-VfR2kkLFssw@public.gmane.org Sender: "dev" On Fri, Nov 14, 2014 at 10:43:39PM +0600, Yerden Zhumabekov wrote: >=20 > 14.11.2014 19:53, Neil Horman =D0=BF=D0=B8=D1=88=D0=B5=D1=82: > > On Fri, Nov 14, 2014 at 05:57:51PM +0600, Yerden Zhumabekov wrote: > >> 14.11.2014 17:33, Neil Horman =D0=BF=D0=B8=D1=88=D0=B5=D1=82: > >>> Not really. That covers the case of applications selecting the has= h function > >>> using the DEFUALT_HASH_FUNC macro, but doesn't nothing for applicat= ions using > >>> the function directly. Test_hash_perf is an example of this, and = ostensibly > >>> because of the behavior without SSE4.2 it defines these huge test t= ables twice > >>> based on the availability of SSE4.2. It would be better if we coul= d allow > >>> applications to use rte_hash_crc regardless, and make the code it u= ses at run > >>> time configurable. > >> I see, then we have a problem here :) > >> > >> Actually, that was one of my concerns when developing these patches.= I > >> looked through the source code of libs and examples and I saw the > >> '#ifdef..#include..#endif'-like appoach while selecting hash functio= n > >> was common. So I organized patches to minimize the impact on API and= not > >> to contradict this approach. > >> > > Thats a reasonable approach, but I really hate the idea of continuing= this need > > to select cpu features at compile time if its not nececcesary. > > > >> If we prefer to change this approach then, I guess, we need to intro= duce > >> broader changes to rte_hash library and change other code which uses= it. > >> If that's what's needed, then it'll take some time for me to rework > >> these patches. > >> > > Well, its possible you'll get lucky. crc is such a common operation,= its > > entirely possible that the gcc intrinsic emits software based crc com= putation if > > the SSE4.2 instructions aren't enabled. I recommend modifying the te= st_hash_crc > > function to use rte_hash_crc with SSE4.2 disabled, and see if you get= a crash. > > If you don't examine the disassembly of your new function and confirm= that > > something reasonable that doesn't use SSE4.2 is emitted. If thats th= e case, > > your patch is fine, and we can focus on how to change the ifdefs in t= he existing > > code, as use of the rte_hash_crc functions should be safe. > > >=20 > Unfortunately, it seems not to be the case. Trying to force compiling a > test program with _mm_crc32_u32 intrinsic on computer with no SSE4.2 > support leads to "Illegal instruction error". So it looks like GCC does > not fall back to crc32 software implementation. >=20 Ok, but crc32 is pretty easy to implement in software. Just appropriate = the calculate_crc32c function from the BSD or Linux kernels and if (unlikely(!support_sse42)) calculate_crc32 operation. Neil > --=20 > Sincerely, >=20 > Yerden Zhumabekov > State Technical Service > Astana, KZ >=20 >=20 >=20