From mboxrd@z Thu Jan 1 00:00:00 1970 Message-ID: <516D0397.5050301@xenomai.org> Date: Tue, 16 Apr 2013 09:53:59 +0200 From: Gilles Chanteperdrix MIME-Version: 1.0 References: <516983AF.2000909@xenomai.org> <516C78FD.1090002@web.de> <516CFC95.4000305@web.de> <516CFE78.1030903@xenomai.org> <516CFEEB.5050704@web.de> In-Reply-To: <516CFEEB.5050704@web.de> Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Subject: Re: [Xenomai] [Xenomai-git] Jan Kiszka : switchtest: Add SSE and AVX check List-Id: Discussions about the Xenomai project List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Jan Kiszka Cc: xenomai@xenomai.org On 04/16/2013 09:34 AM, Jan Kiszka wrote: > On 2013-04-16 09:32, Gilles Chanteperdrix wrote: >> On 04/16/2013 09:24 AM, Jan Kiszka wrote: >> >>> On 2013-04-16 00:02, Jan Kiszka wrote: >>>> On 2013-04-13 18:11, Gilles Chanteperdrix wrote: >>>>> This routine causes a warning from gcc and looks indeed wrong: if the >>>>> "vec" variable is used as an output variable of the inline assembly, it >>>>> should be in the output section of the inline assembly, not the input >>>>> section. >>>>> >>>> >>>> Yes, seems wrong. Will try to look into it the next days. >>> >>> Done, you can find the obvious fix in my for-upstream queue. >> >> >> Hi Jan, >> >> yes the fix is obvious, however the fact that machines with AVX support >> still pass xeno-regression-test is not, have you checked that? > > I've checked that the binary output is unaffected, at least for my > compiler here. Makes sense as there is no other use for the registers > around, and the assembler just blindly put the input variables into the > destination slot of the instructions. No, it does not really make sense: a piece of assembly without any side effects and any outputs is basically useless, so the compiler could very well decide to optimize it out. But if you have checked, OK. -- Gilles.