Attachment 11034 Details for Bug 11380 – diff of assembly code of r300SetupPixelShader() when building with "-O2" and "-O2 -fno-tree-vrp"

diff of assembly code of r300SetupPixelShader() when building with "-O2" and "-O2 -fno-tree-vrp"

r300_state_setuppixelshader.notreevrp.diff (text/x-patch), 4.76 KB, created by Olivier Blin on 2007-08-07 08:24:45 UTC

(hide)

Description:

Filename:

MIME Type:

Creator: Olivier Blin

Created: 2007-08-07 08:24:45 UTC

Size: 4.76 KB

patch

obsolete

>--- /home/blino/vc/mandriva/packages/cooker/mesa/BUILD/Mesa-7.0.1/src/mesa/drivers/dri/r300/r300_state_setuppixelshader.O2.s	2007-08-07 16:19:54.000000000 +0200
>+++ /home/blino/vc/mandriva/packages/cooker/mesa/BUILD/Mesa-7.0.1/src/mesa/drivers/dri/r300/r300_state_setuppixelshader.notreevrp.s	2007-08-07 16:20:37.000000000 +0200
>@@ -17,6 +17,7 @@
> # -DUSE_SSE_ASM -DCOMPILE_R300 -DR200_MERGED=0 -DRADEON_COMMON=1
> # -DRADEON_COMMON_FOR_R300 -mtune=generic -auxbase-strip -g -O -O2 -Wall
> # -Wmissing-prototypes -std=c99 -ffast-math -fverbose-asm -fPIC
>+# -fno-tree-vrp
> # options enabled:  -fPIC -falign-loops -fargument-alias -fbranch-count-reg
> # -fcaller-saves -fcommon -fcprop-registers -fcrossjumping
> # -fcse-follow-jumps -fcse-skip-blocks -fcx-limited-range -fdefer-pop
>@@ -36,7 +37,7 @@
> # -ftree-copyrename -ftree-dce -ftree-dominator-opts -ftree-dse -ftree-fre
> # -ftree-loop-im -ftree-loop-ivcanon -ftree-loop-optimize -ftree-lrs
> # -ftree-pre -ftree-salias -ftree-sink -ftree-sra -ftree-store-ccp
>-# -ftree-store-copy-prop -ftree-ter -ftree-vect-loop-version -ftree-vrp
>+# -ftree-store-copy-prop -ftree-ter -ftree-vect-loop-version
> # -funit-at-a-time -funsafe-math-optimizations -fvar-tracking -fverbose-asm
> # -fzero-initialized-in-bss -m32 -m80387 -m96bit-long-double
> # -maccumulate-outgoing-args -malign-stringops -mfancy-math-387
>@@ -319,24 +320,31 @@
> 	jg	.L68	#,
> .L19:
> 	.loc 1 74 0
>-	movl	-84(%ebp), %edx	#,
>+	movl	-36(%ebp), %eax	# i.233, tmp217
> 	.loc 1 61 0
> 	addl	$4, %edi	#, ivtmp.168
> 	addl	$20, %ecx	#, ivtmp.176
>+	.loc 1 74 0
>+	movl	-84(%ebp), %edx	#,
>+	.loc 1 61 0
> 	addl	$1, -36(%ebp)	#, i.233
> .LVL15:
>+	.loc 1 74 0
>+	sall	$2, %eax	#, tmp217
>+	negl	%eax	# tmp217
>+	.loc 1 61 0
> 	cmpl	$4, -36(%ebp)	#, i.233
> 	.loc 1 74 0
>-	movl	$0, 32(%edx)	#,
>+	movl	$0, 32(%edx,%eax)	#,
> 	.loc 1 61 0
> 	jne	.L69	#,
> .L22:
> 	.loc 1 79 0
> 	movl	3580(%esi), %eax	# <variable>.first_node_has_tex, <variable>.first_node_has_tex
> 	movl	-84(%ebp), %edx	#,
>-	sall	$3, %eax	#, tmp218
>-	orl	3576(%esi), %eax	# <variable>.cur_node, tmp218
>-	movl	%eax, 4(%edx)	# tmp218,
>+	sall	$3, %eax	#, tmp221
>+	orl	3576(%esi), %eax	# <variable>.cur_node, tmp221
>+	movl	%eax, 4(%edx)	# tmp221,
> 	.loc 1 82 0
> 	movl	3668(%esi), %eax	# <variable>.max_temp_idx, <variable>.max_temp_idx
> 	movl	%eax, 8(%edx)	# <variable>.max_temp_idx,
>@@ -382,7 +390,7 @@
> 	.loc 1 92 0
> 	movl	-64(%ebp), %eax	# D.25918,
> 	.loc 1 94 0
>-	movl	%ecx, -44(%ebp)	# <variable>.hw.fpp.cmd, temp.218
>+	movl	%ecx, -44(%ebp)	# <variable>.hw.fpp.cmd, temp.221
> 	.loc 1 92 0
> 	movl	%edx, 4(%edi,%eax)	# float24.239,
> 	.loc 1 94 0
>@@ -403,9 +411,9 @@
> .LBE50:
> .LBE47:
> 	.loc 1 94 0
>-	movl	-44(%ebp), %eax	# temp.218,
>+	movl	-44(%ebp), %eax	# temp.221,
> 	.loc 1 96 0
>-	movl	%ecx, -40(%ebp)	# <variable>.hw.fpp.cmd, temp.219
>+	movl	%ecx, -40(%ebp)	# <variable>.hw.fpp.cmd, temp.222
> 	.loc 1 94 0
> 	movl	%edx, 8(%eax,%edi)	# float24.237,
> 	.loc 1 96 0
>@@ -426,7 +434,7 @@
> .LBE57:
> .LBE54:
> 	.loc 1 96 0
>-	movl	-40(%ebp), %eax	# temp.219,
>+	movl	-40(%ebp), %eax	# temp.222,
> 	.loc 1 98 0
> 	addl	$1, -32(%ebp)	#, i.234
> .LVL22:
>@@ -435,12 +443,12 @@
> 	.loc 1 98 0
> 	movl	-32(%ebp), %edx	# i.234,
> .LVL23:
>-	movl	%ecx, %edi	# <variable>.hw.fpp.cmd, temp.220
>-	movl	3596(%esi,%edx,4), %eax	# <variable>.constant, tmp248
>+	movl	%ecx, %edi	# <variable>.hw.fpp.cmd, temp.223
>+	movl	3596(%esi,%edx,4), %eax	# <variable>.constant, tmp251
> .LBB61:
> .LBB64:
> 	.loc 3 106 0
>-	xorl	%edx, %edx	# float24.222
>+	xorl	%edx, %edx	# float24.225
> .LVL24:
> .LBE64:
> .LBE61:
>@@ -458,9 +466,9 @@
> .LBE67:
> .LBE69:
> 	.loc 1 98 0
>-	movl	-32(%ebp), %eax	# i.234, tmp253
>-	sall	$4, %eax	#, tmp253
>-	movl	%edx, (%eax,%edi)	# float24.222,* temp.220
>+	movl	-32(%ebp), %eax	# i.234, tmp256
>+	sall	$4, %eax	#, tmp256
>+	movl	%edx, (%eax,%edi)	# float24.225,* temp.223
> 	.loc 1 91 0
> 	movl	3664(%esi), %eax	# <variable>.const_nr, D.25945
> 	cmpl	-32(%ebp), %eax	# i.234, D.25945
>@@ -611,13 +619,13 @@
> 	movl	%edx, -16(%ebp)	# D.26085, exponent
> .LVL43:
> 	.loc 3 118 0
>-	sall	$16, %edx	#, float24.222
>+	sall	$16, %edx	#, float24.225
> .LVL44:
>-	orl	%eax, %edx	# u, float24.222
>+	orl	%eax, %edx	# u, float24.225
> 	movl	8(%ebp), %eax	# rmesa,
> .LVL45:
> 	.loc 3 120 0
>-	orl	%ecx, %edx	# float24, float24.222
>+	orl	%ecx, %edx	# float24, float24.225
> 	movl	1732(%eax), %ecx	# <variable>.hw.fpp.cmd, <variable>.hw.fpp.cmd
> 	jmp	.L51	#
> .LVL46:
>@@ -750,16 +758,16 @@
> 	.loc 2 67 0
> 	movl	-80(%ebp), %edx	# cmd,
> .LVL66:
>-	sall	$2, %eax	#, tmp254
>-	movb	%al, %dh	# tmp254,
>+	sall	$2, %eax	#, tmp257
>+	movb	%al, %dh	# tmp257,
> 	.loc 2 68 0
>-	movzwl	%dx, %eax	#, tmp257
>+	movzwl	%dx, %eax	#, tmp260
> 	.loc 2 69 0
>-	orl	$1275068416, %eax	#, tmp257
>+	orl	$1275068416, %eax	#, tmp260
> .LBE76:
> .LBE75:
> 	.loc 1 101 0
>-	movl	%eax, (%ecx)	# tmp257,* <variable>.hw.fpp.cmd
>+	movl	%eax, (%ecx)	# tmp260,* <variable>.hw.fpp.cmd
> 	.loc 1 103 0
> 	addl	$92, %esp	#,
> 	popl	%ebx	#

Actions: View

Attachments on bug 11380: 10459 | 11032 | 11033 | 11034

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.