From owner-svn-src-all@freebsd.org Tue Sep 3 14:07:40 2019 Return-Path: Delivered-To: svn-src-all@mailman.nyi.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2610:1c1:1:606c::19:1]) by mailman.nyi.freebsd.org (Postfix) with ESMTP id C9158DD074; Tue, 3 Sep 2019 14:06:51 +0000 (UTC) (envelope-from yuripv@freebsd.org) Received: from freefall.freebsd.org (freefall.freebsd.org [IPv6:2610:1c1:1:6074::16:84]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) server-signature RSA-PSS (4096 bits) client-signature RSA-PSS (4096 bits) client-digest SHA256) (Client CN "freefall.freebsd.org", Issuer "Let's Encrypt Authority X3" (verified OK)) by mx1.freebsd.org (Postfix) with ESMTPS id 46N7zv1Ysrz4PyM; Tue, 3 Sep 2019 14:06:51 +0000 (UTC) (envelope-from yuripv@freebsd.org) Received: by freefall.freebsd.org (Postfix, from userid 1452) id 193A11AC79; Tue, 3 Sep 2019 14:06:19 +0000 (UTC) X-Original-To: yuripv@localmail.freebsd.org Delivered-To: yuripv@localmail.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2610:1c1:1:606c::19:1]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits)) (Client CN "mx1.freebsd.org", Issuer "Let's Encrypt Authority X3" (verified OK)) by freefall.freebsd.org (Postfix) with ESMTPS id 7704C1632B; Sun, 14 Apr 2019 13:37:54 +0000 (UTC) (envelope-from owner-src-committers@freebsd.org) Received: from freefall.freebsd.org (freefall.freebsd.org [96.47.72.132]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) server-signature RSA-PSS (4096 bits) client-signature RSA-PSS (4096 bits) client-digest SHA256) (Client CN "freefall.freebsd.org", Issuer "Let's Encrypt Authority X3" (verified OK)) by mx1.freebsd.org (Postfix) with ESMTPS id 285246AFED; Sun, 14 Apr 2019 13:37:54 +0000 (UTC) (envelope-from owner-src-committers@freebsd.org) Received: by freefall.freebsd.org (Postfix, from userid 538) id 0DA9016329; Sun, 14 Apr 2019 13:37:54 +0000 (UTC) Delivered-To: src-committers@localmail.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2610:1c1:1:606c::19:1]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits)) (Client CN "mx1.freebsd.org", Issuer "Let's Encrypt Authority X3" (verified OK)) by freefall.freebsd.org (Postfix) with ESMTPS id 74D0B16327 for ; Sun, 14 Apr 2019 13:37:51 +0000 (UTC) (envelope-from bde@FreeBSD.org) Received: from mxrelay.nyi.freebsd.org (mxrelay.nyi.freebsd.org [IPv6:2610:1c1:1:606c::19:3]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) server-signature RSA-PSS (4096 bits) client-signature RSA-PSS (4096 bits) client-digest SHA256) (Client CN "mxrelay.nyi.freebsd.org", Issuer "Let's Encrypt Authority X3" (verified OK)) by mx1.freebsd.org (Postfix) with ESMTPS id 4369C6AFEB; Sun, 14 Apr 2019 13:37:51 +0000 (UTC) (envelope-from bde@FreeBSD.org) Received: from repo.freebsd.org (repo.freebsd.org [IPv6:2610:1c1:1:6068::e6a:0]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (Client did not present a certificate) by mxrelay.nyi.freebsd.org (Postfix) with ESMTPS id 1C342212B9; Sun, 14 Apr 2019 13:37:51 +0000 (UTC) (envelope-from bde@FreeBSD.org) Received: from repo.freebsd.org ([127.0.1.37]) by repo.freebsd.org (8.15.2/8.15.2) with ESMTP id x3EDbokg084986; Sun, 14 Apr 2019 13:37:50 GMT (envelope-from bde@FreeBSD.org) Received: (from bde@localhost) by repo.freebsd.org (8.15.2/8.15.2/Submit) id x3EDboMO084985; Sun, 14 Apr 2019 13:37:50 GMT (envelope-from bde@FreeBSD.org) Message-Id: <201904141337.x3EDboMO084985@repo.freebsd.org> X-Authentication-Warning: repo.freebsd.org: bde set sender to bde@FreeBSD.org using -f From: Bruce Evans To: src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-head@freebsd.org Subject: svn commit: r346215 - head/lib/libvgl X-SVN-Group: head X-SVN-Commit-Author: bde X-SVN-Commit-Paths: head/lib/libvgl X-SVN-Commit-Revision: 346215 X-SVN-Commit-Repository: base MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Precedence: bulk X-Loop: FreeBSD.org Sender: owner-src-committers@freebsd.org X-Rspamd-Queue-Id: 285246AFED X-Spamd-Bar: -- Authentication-Results: mx1.freebsd.org X-Spamd-Result: default: False [-2.98 / 15.00]; local_wl_from(0.00)[freebsd.org]; NEURAL_HAM_MEDIUM(-1.00)[-0.999,0]; NEURAL_HAM_SHORT(-0.98)[-0.983,0]; ASN(0.00)[asn:11403, ipnet:96.47.64.0/20, country:US]; NEURAL_HAM_LONG(-1.00)[-1.000,0] Status: O X-BeenThere: svn-src-all@freebsd.org X-Mailman-Version: 2.1.29 List-Id: "SVN commit messages for the entire src tree \(except for " user" and " projects" \)" List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Date: Tue, 03 Sep 2019 14:07:40 -0000 X-Original-Date: Sun, 14 Apr 2019 13:37:50 +0000 (UTC) X-List-Received-Date: Tue, 03 Sep 2019 14:07:40 -0000 Author: bde Date: Sun Apr 14 13:37:50 2019 New Revision: 346215 URL: https://svnweb.freebsd.org/changeset/base/346215 Log: For writing and reading single pixels, avoid some pessimizations for depths > 8. Add some smaller optimizations for these depths. Use a more generic method for all depths >= 8, although this gives tiny pessimizations for these depths. For clearing the whole frame buffer, avoid the same pessimizations for depths > 8. Add some larger optimizations for these depths. Use an even more generic method for all depths >= 8 to give the optimizations for depths > 8 and a tiny pessimization for depth 8. The main pessimization was that old versions of bcopy() copy 1 byte at a time for all trailing bytes. (i386 still does this. amd64 now pessimizzes large sizes instead of small ones if the CPU supports ERMS. dev/fb gets this wrong by mostly not using the bcopy() family or the technically correct bus space functions but by mostly copying 2 bytes at a time using an unoptimized loop without even volatile declarations to prevent the compiler rewriting it.) The sizes here are 1, 2, 3 or 4 bytes, so depths 9-16 were up to twice as slow as necessary and depths 17-24 were up to 3 times slower than necessary. Fix this (except depths 17-24 are still up to 2 times slower than necessary) by using (builtin) memcpy() instead of bcopy() and reorganizing so that the complier can see the small constant sizes. Reduce special cases while reorganizing although this is slightly slower than adding special cases. The compiler inlining (and even -O2 vs -O0) makes little difference compared with reducing the number of accesses except on modern hardware it gives a small improvement. Clearing was also pessimized mainly by the extra accesses. Fix it quite differently by creating a MEMBUF containing 1 line (in fast memory using a slow method) and copying this. This is only slightly slower than reducing everything to efficient memset()s and bcopy()s, but simpler, especially for the segmented case. This works for planar modes too, but don't use it then since the old method was actually optimal for planar modes (it works by moving the slow i/o instructions out of inner loops), while for direct modes the slow instructions were all in the invisible inner loop in bcopy(). Use htole32() and le32toh() and some type puns instead of unoptimized functions for converting colors. This optimization is mostly in the noise. libvgl is only supported on x86, so it could hard-code the assumption that the byte order is le32, but the old conversion functions didn't hard-code this. Modified: head/lib/libvgl/simple.c Modified: head/lib/libvgl/simple.c ============================================================================== --- head/lib/libvgl/simple.c Sun Apr 14 13:17:40 2019 (r346214) +++ head/lib/libvgl/simple.c Sun Apr 14 13:37:50 2019 (r346215) @@ -33,6 +33,7 @@ __FBSDID("$FreeBSD$"); #include #include +#include #include "vgl.h" static byte VGLSavePaletteRed[256]; @@ -44,96 +45,44 @@ static byte VGLSavePaletteBlue[256]; #define min(x, y) (((x) < (y)) ? (x) : (y)) #define max(x, y) (((x) > (y)) ? (x) : (y)) -static void -color2mem(u_long color, byte *b, int len) -{ - switch (len) { - case 4: - b[3] = (color >> 24) & 0xff; - /* fallthrough */ - case 3: - b[2] = (color >> 16) & 0xff; - /* fallthrough */ - case 2: - b[1] = (color >> 8) & 0xff; - /* fallthrough */ - case 1: - default: - b[0] = color & 0xff; - break; - } - - return; -} - -static u_long -mem2color(byte *b, int len) -{ - u_long color = 0; - - switch (len) { - case 4: - color |= (b[3] & 0xff) << 24; - /* fallthrough */ - case 3: - color |= (b[2] & 0xff) << 16; - /* fallthrough */ - case 2: - color |= (b[1] & 0xff) << 8; - /* fallthrough */ - case 1: - default: - color |= (b[0] & 0xff); - break; - } - - return color; -} - void VGLSetXY(VGLBitmap *object, int x, int y, u_long color) { int offset; - byte b[4]; VGLCheckSwitch(); if (x>=0 && xVXsize && y>=0 && yVYsize) { if (object->Type == MEMBUF || !VGLMouseFreeze(x, y, 1, 1, 0x80000000 | color)) { + offset = (y * object->VXsize + x) * object->PixelBytes; switch (object->Type) { + case VIDBUF8S: + case VIDBUF16S: + case VIDBUF24S: + case VIDBUF32S: + offset = VGLSetSegment(offset); + /* FALLTHROUGH */ case MEMBUF: - switch (object->PixelBytes) { - case 2: - goto vidbuf16; - case 3: - goto vidbuf24; - case 4: - goto vidbuf32; - } - /* fallthrough */ case VIDBUF8: - object->Bitmap[y*object->VXsize+x]=((byte)color); - break; - case VIDBUF8S: - object->Bitmap[VGLSetSegment(y*object->VXsize+x)]=((byte)color); - break; case VIDBUF16: -vidbuf16: case VIDBUF24: -vidbuf24: case VIDBUF32: -vidbuf32: - color2mem(color, b, object->PixelBytes); - bcopy(b, &object->Bitmap[(y*object->VXsize+x) * object->PixelBytes], - object->PixelBytes); + color = htole32(color); + switch (object->PixelBytes) { + case 1: + memcpy(&object->Bitmap[offset], &color, 1); + break; + case 2: + memcpy(&object->Bitmap[offset], &color, 2); + break; + case 3: + memcpy(&object->Bitmap[offset], &color, 3); + break; + case 4: + memcpy(&object->Bitmap[offset], &color, 4); + break; + } break; - case VIDBUF16S: - case VIDBUF24S: - case VIDBUF32S: - color2mem(color, b, object->PixelBytes); - offset = VGLSetSegment((y*object->VXsize+x) * object->PixelBytes); - bcopy(b, &object->Bitmap[offset], object->PixelBytes); - break; case VIDBUF8X: outb(0x3c4, 0x02); outb(0x3c5, 0x01 << (x&0x3)); @@ -161,42 +110,38 @@ static u_long __VGLGetXY(VGLBitmap *object, int x, int y) { int offset; - byte b[4]; int i; u_long color; byte mask; + offset = (y * object->VXsize + x) * object->PixelBytes; switch (object->Type) { + case VIDBUF8S: + case VIDBUF16S: + case VIDBUF24S: + case VIDBUF32S: + offset = VGLSetSegment(offset); + /* FALLTHROUGH */ case MEMBUF: - switch (object->PixelBytes) { - case 2: - goto vidbuf16; - case 3: - goto vidbuf24; - case 4: - goto vidbuf32; - } - /* fallthrough */ case VIDBUF8: - return object->Bitmap[((y*object->VXsize)+x)]; - case VIDBUF8S: - return object->Bitmap[VGLSetSegment(y*object->VXsize+x)]; case VIDBUF16: -vidbuf16: case VIDBUF24: -vidbuf24: case VIDBUF32: -vidbuf32: - bcopy(&object->Bitmap[(y*object->VXsize+x) * object->PixelBytes], - b, object->PixelBytes); - return (mem2color(b, object->PixelBytes)); - case VIDBUF16S: - case VIDBUF24S: - case VIDBUF32S: - offset = VGLSetSegment((y*object->VXsize+x) * object->PixelBytes); - bcopy(&object->Bitmap[offset], b, object->PixelBytes); - - return (mem2color(b, object->PixelBytes)); + switch (object->PixelBytes) { + case 1: + memcpy(&color, &object->Bitmap[offset], 1); + return le32toh(color) & 0xff; + case 2: + memcpy(&color, &object->Bitmap[offset], 2); + return le32toh(color) & 0xffff; + case 3: + memcpy(&color, &object->Bitmap[offset], 3); + return le32toh(color) & 0xffffff; + case 4: + memcpy(&color, &object->Bitmap[offset], 4); + return le32toh(color); + } + break; case VIDBUF8X: outb(0x3ce, 0x04); outb(0x3cf, x & 0x3); return object->Bitmap[(unsigned)(VGLAdpInfo.va_line_width*y)+(x/4)]; @@ -539,63 +484,38 @@ VGLFilledEllipse(VGLBitmap *object, int xc, int yc, in void VGLClear(VGLBitmap *object, u_long color) { + VGLBitmap src; int offset; int len; - int i, total = 0; - byte b[4]; + int i; VGLCheckSwitch(); if (object->Type != MEMBUF) VGLMouseFreeze(0, 0, object->Xsize, object->Ysize, color); switch (object->Type) { case MEMBUF: - switch (object->PixelBytes) { - case 2: - goto vidbuf16; - case 3: - goto vidbuf24; - case 4: - goto vidbuf32; - } - /* fallthrough */ case VIDBUF8: - memset(object->Bitmap, (byte)color, object->VXsize*object->VYsize); - break; - case VIDBUF8S: - for (offset = 0; offset < object->VXsize*object->VYsize; ) { - VGLSetSegment(offset); - len = min(object->VXsize*object->VYsize - offset, - VGLAdpInfo.va_window_size); - memset(object->Bitmap, (byte)color, len); - offset += len; - } - break; case VIDBUF16: -vidbuf16: - case VIDBUF24: -vidbuf24: - case VIDBUF32: -vidbuf32: - color2mem(color, b, object->PixelBytes); - total = object->VXsize*object->VYsize*object->PixelBytes; - for (i = 0; i < total; i += object->PixelBytes) - bcopy(b, object->Bitmap + i, object->PixelBytes); - break; - case VIDBUF16S: + case VIDBUF24: case VIDBUF24S: + case VIDBUF32: case VIDBUF32S: - color2mem(color, b, object->PixelBytes); - total = object->VXsize*object->VYsize*object->PixelBytes; - for (offset = 0; offset < total; ) { - VGLSetSegment(offset); - len = min(total - offset, VGLAdpInfo.va_window_size); - for (i = 0; i < len; i += object->PixelBytes) - bcopy(b, object->Bitmap + (offset + i) % VGLAdpInfo.va_window_size, - object->PixelBytes); - offset += len; - } + src.Type = MEMBUF; + src.Xsize = object->Xsize; + src.VXsize = object->VXsize; + src.Ysize = 1; + src.VYsize = 1; + src.Xorigin = 0; + src.Yorigin = 0; + src.Bitmap = alloca(object->VXsize * object->PixelBytes); + src.PixelBytes = object->PixelBytes; + color = htole32(color); + for (i = 0; i < object->VXsize; i++) + bcopy(&color, src.Bitmap + i * object->PixelBytes, object->PixelBytes); + for (i = 0; i < object->VYsize; i++) + __VGLBitmapCopy(&src, 0, 0, object, 0, i, object->VYsize, 1); break; case VIDBUF8X: