From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-6.8 required=3.0 tests=DKIM_SIGNED,DKIM_VALID, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_PATCH,MAILING_LIST_MULTI,SIGNED_OFF_BY, SPF_HELO_NONE,SPF_PASS autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 9BA48C54FD0 for ; Mon, 27 Apr 2020 15:13:31 +0000 (UTC) Received: from alsa0.perex.cz (alsa0.perex.cz [77.48.224.243]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 24E262054F for ; Mon, 27 Apr 2020 15:13:31 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=pass (1024-bit key) header.d=alsa-project.org header.i=@alsa-project.org header.b="r+xn8wM3" DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 24E262054F Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=suse.de Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=alsa-devel-bounces@alsa-project.org Received: from alsa1.perex.cz (alsa1.perex.cz [207.180.221.201]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by alsa0.perex.cz (Postfix) with ESMTPS id 7AE0C1673; Mon, 27 Apr 2020 17:12:39 +0200 (CEST) DKIM-Filter: OpenDKIM Filter v2.11.0 alsa0.perex.cz 7AE0C1673 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=alsa-project.org; s=default; t=1588000409; bh=lO245DByOcspDbBLGi9jOiUIPSScoFIYi+WLnYHJ0PI=; h=Date:From:To:Subject:In-Reply-To:References:Cc:List-Id: List-Unsubscribe:List-Archive:List-Post:List-Help:List-Subscribe: From; b=r+xn8wM3W/kee1jxukSCQyZu9mW58Q8WOMoA2x+RJYPMCNUi8UKLYPkfA2NV/OSuz XNMFHej4uJ/QC7rhqnJROEJa3NdjJLzsgwWU2QRN+MuCvB/+RMcHuvHnhoME7jGTH0 z5NPx0Zc3+k7eFreccdFuQBBnswi1aCdu/IfHXJY= Received: from alsa1.perex.cz (localhost.localdomain [127.0.0.1]) by alsa1.perex.cz (Postfix) with ESMTP id 1ECC6F8010A; Mon, 27 Apr 2020 17:12:39 +0200 (CEST) Received: by alsa1.perex.cz (Postfix, from userid 50401) id C8F04F80113; Mon, 27 Apr 2020 17:12:36 +0200 (CEST) Received: from mx2.suse.de (mx2.suse.de [195.135.220.15]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by alsa1.perex.cz (Postfix) with ESMTPS id 7492AF80113 for ; Mon, 27 Apr 2020 17:12:29 +0200 (CEST) DKIM-Filter: OpenDKIM Filter v2.11.0 alsa1.perex.cz 7492AF80113 X-Virus-Scanned: by amavisd-new at test-mx.suse.de Received: from relay2.suse.de (unknown [195.135.220.254]) by mx2.suse.de (Postfix) with ESMTP id 7AB9BABCF; Mon, 27 Apr 2020 15:12:28 +0000 (UTC) Date: Mon, 27 Apr 2020 17:12:28 +0200 Message-ID: From: Takashi Iwai To: =?UTF-8?B?RlLDiUTDiVJJQw==?= RECOULES Subject: Re: [PATCH] [inline assembly] fix pcm_dmix_i386.h assembly chunk interfaces In-Reply-To: <1871992915.6898305.1587998132827.JavaMail.zimbra@univ-grenoble-alpes.fr> References: <20200427073604.26662-1-frederic.recoules@univ-grenoble-alpes.fr> <1871992915.6898305.1587998132827.JavaMail.zimbra@univ-grenoble-alpes.fr> User-Agent: Wanderlust/2.15.9 (Almost Unreal) SEMI/1.14.6 (Maruoka) FLIM/1.14.9 (=?UTF-8?B?R29qxY0=?=) APEL/10.8 Emacs/25.3 (x86_64-suse-linux-gnu) MULE/6.0 (HANACHIRUSATO) MIME-Version: 1.0 (generated by SEMI 1.14.6 - "Maruoka") Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Cc: alsa-devel@alsa-project.org X-BeenThere: alsa-devel@alsa-project.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: "Alsa-devel mailing list for ALSA developers - http://www.alsa-project.org" List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: alsa-devel-bounces@alsa-project.org Sender: "Alsa-devel" Please don't drop Cc to ML. And about the comments: On Mon, 27 Apr 2020 16:35:32 +0200, FRÉDÉRIC RECOULES wrote: > > I wrongly assumed that the option -mmmx was passed when compiling (re) > mix_areas_16_mmx. > I know how to fix it (inspired of ffmpeg): > - the configure test if mmx is supported by the compiler option and set a > macro HAVE_MMX accordingly (maybe something already exist?). Actually the mmx support isn't about whether the compiler supports it or not. With inline asm, the MMX-enabled code is always included in alsa-lib as well as the code without MMX, then which one to be executed is dynamically switched at runtime by probing the CPU capability. thanks, Takashi > - declare a macro MMX_CLOBBERS(list) that will output the list when HAVE_MMX > is true and nothing otherwise. > I will resubmit a patch soon. > > Regards, > Frédéric Recoules > > ------------------------------------------------------------------------------ > De: "tiwai" > À: "frederic recoules" > Cc: "alsa-devel" , "frederic recoules" > > Envoyé: Lundi 27 Avril 2020 15:47:02 > Objet: Re: [PATCH] [inline assembly] fix pcm_dmix_i386.h assembly chunk > interfaces > > On Mon, 27 Apr 2020 09:36:04 +0200, > frederic.recoules@univ-grenoble-alpes.fr wrote: > > > > From: Frédéric Recoules > > > > Main changes are: > > - move 'size' and 'old_ebx' to the output list since they are clobbered > > - add the "memory" keyword since input pointers are dereferenced > > - add mmx registers in the clobber list and add an initialization for mm1 > > - add ebx in clobbers via a set of macro when GCC is newer than 5.0 > > (it will work for other compilers or non-PIC mode too) > > > > Minor changes are: > > - keep consistent the token numbering in the template > > - remove the manual save/restore ebx when it is in the clobber list > > - allows 'dst_step', 'src_step' and 'sum_step' to be given by immediates > > - allows 'size' to be given by register (e.g. ebp) > > - add "cc" keyword since the eflag register is clobbered > > > > Signed-off-by: Frédéric Recoules > > When I apply this and build for i386 with gcc9, I got the following > error: > pcm_dmix_i386.h: In function 'mix_areas_16_mmx': > pcm_dmix_i386.h:180:2: error: unknown register name 'mm1' in 'asm' > 180 | __asm__ __volatile__ ( > | ^~~~~~~ > pcm_dmix_i386.h:180:2: error: unknown register name 'mm0' in 'asm' > In file included from pcm_dmix_i386.c:31, > from pcm_dmix.c:144: > pcm_dmix_i386.h: In function 'remix_areas_16_mmx': > pcm_dmix_i386.h:180:2: error: unknown register name 'mm1' in 'asm' > 180 | __asm__ __volatile__ ( > | ^~~~~~~ > .... > > Could you check those errors? > > thanks, > > Takashi > > > --- > > src/pcm/pcm_dmix_i386.h | 168 ++++++++++++++++++++++------------------ > > 1 file changed, 93 insertions(+), 75 deletions(-) > > > > diff --git a/src/pcm/pcm_dmix_i386.h b/src/pcm/pcm_dmix_i386.h > > index 2778cb1d..af2f4630 100644 > > --- a/src/pcm/pcm_dmix_i386.h > > +++ b/src/pcm/pcm_dmix_i386.h > > @@ -26,6 +26,13 @@ > > * > > */ > > > > +#define COMMA , > > +#if __GNUC__ < 5 && defined(__PIC__) > > +# define GCC_PIC_SWITCH(before,after) before > > +#else > > +# define GCC_PIC_SWITCH(before,after) after > > +#endif > > + > > /* > > * for plain i386 > > */ > > @@ -47,13 +54,14 @@ static void MIX_AREAS_16(unsigned int size, > > __asm__ __volatile__ ( > > "\n" > > > > - "\tmovl %%ebx, %7\n" /* ebx is GOT pointer (-fPIC) * > / > > + /* ebx is GOT pointer (-fPIC) */ > > + GCC_PIC_SWITCH("\tmovl %%ebx, %1\n",) > > /* > > * initialization, load ESI, EDI, EBX registers > > */ > > - "\tmovl %1, %%edi\n" > > - "\tmovl %2, %%esi\n" > > - "\tmovl %3, %%ebx\n" > > + "\tmovl %2, %%edi\n" > > + "\tmovl %3, %%esi\n" > > + "\tmovl %4, %%ebx\n" > > "\tcmpl $0, %0\n" > > "\tjnz 2f\n" > > "\tjmp 7f\n" > > @@ -64,9 +72,9 @@ static void MIX_AREAS_16(unsigned int size, > > */ > > "\t.p2align 4,,15\n" > > "1:" > > - "\tadd %4, %%edi\n" > > - "\tadd %5, %%esi\n" > > - "\tadd %6, %%ebx\n" > > + "\tadd %5, %%edi\n" > > + "\tadd %6, %%esi\n" > > + "\tadd %7, %%ebx\n" > > > > /* > > * sample = *src; > > @@ -138,15 +146,16 @@ static void MIX_AREAS_16(unsigned int size, > > "\tjnz 4b\n" > > "\tdecl %0\n" > > "\tjnz 1b\n" > > - > > - "7:" > > - "\tmovl %7, %%ebx\n" /* ebx is GOT pointer (-fPIC) * > / > > > > - : /* no output regs */ > > - : "m" (size), "m" (dst), "m" (src), > > - "m" (sum), "m" (dst_step), "m" (src_step), > > - "m" (sum_step), "m" (old_ebx) > > - : "esi", "edi", "edx", "ecx", "eax" > > + "7:" > > + /* ebx is GOT pointer (-fPIC) */ > > + GCC_PIC_SWITCH("\tmovl %1, %%ebx\n",) > > + > > + : "+&rm" (size), GCC_PIC_SWITCH("=m","=X") (old_ebx) > > + : "m" (dst), "m" (src), "m" (sum), > > + "im" (dst_step), "im" (src_step), "im" (sum_step) > > + : "esi", "edi", "edx", "ecx", GCC_PIC_SWITCH(,"ebx"COMMA) > "eax", > > + "memory", "cc" > > ); > > } > > > > @@ -171,22 +180,24 @@ static void MIX_AREAS_16_MMX(unsigned int size, > > __asm__ __volatile__ ( > > "\n" > > > > - "\tmovl %%ebx, %7\n" /* ebx is GOT pointer (-fPIC) * > / > > + /* ebx is GOT pointer (-fPIC) */ > > + GCC_PIC_SWITCH("\tmovl %%ebx, %1\n",) > > /* > > - * initialization, load ESI, EDI, EBX registers > > + * initialization, load ESI, EDI, EBX registers, clear MM1 > > */ > > - "\tmovl %1, %%edi\n" > > - "\tmovl %2, %%esi\n" > > - "\tmovl %3, %%ebx\n" > > + "\tpxor %%mm1, %%mm1\n" > > + "\tmovl %2, %%edi\n" > > + "\tmovl %3, %%esi\n" > > + "\tmovl %4, %%ebx\n" > > "\tcmpl $0, %0\n" > > "\tjnz 2f\n" > > "\tjmp 5f\n" > > > > "\t.p2align 4,,15\n" > > "1:" > > - "\tadd %4, %%edi\n" > > - "\tadd %5, %%esi\n" > > - "\tadd %6, %%ebx\n" > > + "\tadd %5, %%edi\n" > > + "\tadd %6, %%esi\n" > > + "\tadd %7, %%ebx\n" > > > > "2:" > > /* > > @@ -230,13 +241,14 @@ static void MIX_AREAS_16_MMX(unsigned int size, > > "\tjnz 1b\n" > > "\temms\n" > > "5:" > > - "\tmovl %7, %%ebx\n" /* ebx is GOT pointer (-fPIC) * > / > > - > > - : /* no output regs */ > > - : "m" (size), "m" (dst), "m" (src), > > - "m" (sum), "m" (dst_step), "m" (src_step), > > - "m" (sum_step), "m" (old_ebx) > > - : "esi", "edi", "edx", "ecx", "eax" > > + /* ebx is GOT pointer (-fPIC) */ > > + GCC_PIC_SWITCH("\tmovl %1, %%ebx\n",) > > + > > + : "+&rm" (size), GCC_PIC_SWITCH("=m","=X") (old_ebx) > > + : "m" (dst), "m" (src), "m" (sum), > > + "im" (dst_step), "im" (src_step), "im" (sum_step) > > + : "esi", "edi", "edx", "ecx", GCC_PIC_SWITCH(,"ebx"COMMA) > "eax", > > + "mm0", "mm1", "memory", "cc" > > ); > > } > > > > @@ -261,13 +273,14 @@ static void MIX_AREAS_32(unsigned int size, > > __asm__ __volatile__ ( > > "\n" > > > > - "\tmovl %%ebx, %7\n" /* ebx is GOT pointer (-fPIC) * > / > > + /* ebx is GOT pointer (-fPIC) */ > > + GCC_PIC_SWITCH("\tmovl %%ebx, %1\n",) > > /* > > * initialization, load ESI, EDI, EBX registers > > */ > > - "\tmovl %1, %%edi\n" > > - "\tmovl %2, %%esi\n" > > - "\tmovl %3, %%ebx\n" > > + "\tmovl %2, %%edi\n" > > + "\tmovl %3, %%esi\n" > > + "\tmovl %4, %%ebx\n" > > "\tcmpl $0, %0\n" > > "\tjnz 1f\n" > > "\tjmp 6f\n" > > @@ -337,19 +350,20 @@ static void MIX_AREAS_32(unsigned int size, > > */ > > "\tdecl %0\n" > > "\tjz 6f\n" > > - "\tadd %4, %%edi\n" > > - "\tadd %5, %%esi\n" > > - "\tadd %6, %%ebx\n" > > + "\tadd %5, %%edi\n" > > + "\tadd %6, %%esi\n" > > + "\tadd %7, %%ebx\n" > > "\tjmp 1b\n" > > - > > - "6:" > > - "\tmovl %7, %%ebx\n" /* ebx is GOT pointer (-fPIC) * > / > > > > - : /* no output regs */ > > - : "m" (size), "m" (dst), "m" (src), > > - "m" (sum), "m" (dst_step), "m" (src_step), > > - "m" (sum_step), "m" (old_ebx) > > - : "esi", "edi", "edx", "ecx", "eax" > > + "6:" > > + /* ebx is GOT pointer (-fPIC) */ > > + GCC_PIC_SWITCH("\tmovl %1, %%ebx\n",) > > + > > + : "+&rm" (size), GCC_PIC_SWITCH("=m","=X") (old_ebx) > > + : "m" (dst), "m" (src), "m" (sum), > > + "im" (dst_step), "im" (src_step), "im" (sum_step) > > + : "esi", "edi", "edx", "ecx", GCC_PIC_SWITCH(,"ebx"COMMA) > "eax", > > + "memory", "cc" > > ); > > } > > > > @@ -374,13 +388,14 @@ static void MIX_AREAS_24(unsigned int size, > > __asm__ __volatile__ ( > > "\n" > > > > - "\tmovl %%ebx, %7\n" /* ebx is GOT pointer (-fPIC) * > / > > + /* ebx is GOT pointer (-fPIC) */ > > + GCC_PIC_SWITCH("\tmovl %%ebx, %1\n",) > > /* > > * initialization, load ESI, EDI, EBX registers > > */ > > - "\tmovl %1, %%edi\n" > > - "\tmovl %2, %%esi\n" > > - "\tmovl %3, %%ebx\n" > > + "\tmovl %2, %%edi\n" > > + "\tmovl %3, %%esi\n" > > + "\tmovl %4, %%ebx\n" > > "\tcmpl $0, %0\n" > > "\tjnz 1f\n" > > "\tjmp 6f\n" > > @@ -443,19 +458,20 @@ static void MIX_AREAS_24(unsigned int size, > > */ > > "\tdecl %0\n" > > "\tjz 6f\n" > > - "\tadd %4, %%edi\n" > > - "\tadd %5, %%esi\n" > > - "\tadd %6, %%ebx\n" > > + "\tadd %5, %%edi\n" > > + "\tadd %6, %%esi\n" > > + "\tadd %7, %%ebx\n" > > "\tjmp 1b\n" > > - > > - "6:" > > - "\tmovl %7, %%ebx\n" /* ebx is GOT pointer (-fPIC) * > / > > > > - : /* no output regs */ > > - : "m" (size), "m" (dst), "m" (src), > > - "m" (sum), "m" (dst_step), "m" (src_step), > > - "m" (sum_step), "m" (old_ebx) > > - : "esi", "edi", "edx", "ecx", "eax" > > + "6:" > > + /* ebx is GOT pointer (-fPIC) */ > > + GCC_PIC_SWITCH("\tmovl %1, %%ebx\n",) > > + > > + : "+&rm" (size), GCC_PIC_SWITCH("=m","=X") (old_ebx) > > + : "m" (dst), "m" (src), "m" (sum), > > + "im" (dst_step), "im" (src_step), "im" (sum_step) > > + : "esi", "edi", "edx", "ecx", GCC_PIC_SWITCH(,"ebx"COMMA) > "eax", > > + "memory", "cc" > > ); > > } > > > > @@ -480,13 +496,14 @@ static void MIX_AREAS_24_CMOV(unsigned int size, > > __asm__ __volatile__ ( > > "\n" > > > > - "\tmovl %%ebx, %7\n" /* ebx is GOT pointer (-fPIC) * > / > > + /* ebx is GOT pointer (-fPIC) */ > > + GCC_PIC_SWITCH("\tmovl %%ebx, %1\n",) > > /* > > * initialization, load ESI, EDI, EBX registers > > */ > > - "\tmovl %1, %%edi\n" > > - "\tmovl %2, %%esi\n" > > - "\tmovl %3, %%ebx\n" > > + "\tmovl %2, %%edi\n" > > + "\tmovl %3, %%esi\n" > > + "\tmovl %4, %%ebx\n" > > "\tcmpl $0, %0\n" > > "\tjz 6f\n" > > > > @@ -541,19 +558,20 @@ static void MIX_AREAS_24_CMOV(unsigned int size, > > /* > > * while (size-- > 0) > > */ > > - "\tadd %4, %%edi\n" > > - "\tadd %5, %%esi\n" > > - "\tadd %6, %%ebx\n" > > + "\tadd %5, %%edi\n" > > + "\tadd %6, %%esi\n" > > + "\tadd %7, %%ebx\n" > > "\tdecl %0\n" > > "\tjnz 1b\n" > > - > > - "6:" > > - "\tmovl %7, %%ebx\n" /* ebx is GOT pointer (-fPIC) * > / > > > > - : /* no output regs */ > > - : "m" (size), "m" (dst), "m" (src), > > - "m" (sum), "m" (dst_step), "m" (src_step), > > - "m" (sum_step), "m" (old_ebx) > > - : "esi", "edi", "edx", "ecx", "eax" > > + "6:" > > + /* ebx is GOT pointer (-fPIC) */ > > + GCC_PIC_SWITCH("\tmovl %1, %%ebx\n",) > > + > > + : "+&rm" (size), GCC_PIC_SWITCH("=m","=X") (old_ebx) > > + : "m" (dst), "m" (src), "m" (sum), > > + "im" (dst_step), "im" (src_step), "im" (sum_step) > > + : "esi", "edi", "edx", "ecx", GCC_PIC_SWITCH(,"ebx"COMMA) > "eax", > > + "memory", "cc" > > ); > > } > > -- > > 2.17.1 > > > >