SOURCES: smpeg-fPIC.patch (NEW) - fix PIC issue.
pluto
pluto at pld-linux.org
Fri Jan 20 21:14:25 CET 2006
Author: pluto Date: Fri Jan 20 20:14:24 2006 GMT
Module: SOURCES Tag: HEAD
---- Log message:
- fix PIC issue.
---- Files affected:
SOURCES:
smpeg-fPIC.patch (NONE -> 1.1) (NEW)
---- Diffs:
================================================================
Index: SOURCES/smpeg-fPIC.patch
diff -u /dev/null SOURCES/smpeg-fPIC.patch:1.1
--- /dev/null Fri Jan 20 21:14:24 2006
+++ SOURCES/smpeg-fPIC.patch Fri Jan 20 21:14:19 2006
@@ -0,0 +1,308 @@
+Fix PIC issues in mmx routines
+
+--- video/mmxflags_asm.S
++++ video/mmxflags_asm.S
+@@ -1,11 +1,6 @@
+
+ #if defined(i386) && defined(USE_MMX)
+
+-.data
+- .align 16
+- .type flags, at object
+-flags: .long 0
+-
+ .text
+ .align 4
+ .globl cpu_flags
+@@ -40,16 +35,13 @@ cpu_flags:
+ xorl %ecx,%eax
+ je cpu_flags.L1
+
+- pusha
++ pushl %ebx
+
+ movl $1,%eax
+ cpuid
++ movl %edx,%eax
+
+- movl %edx,flags
+-
+- popa
+-
+- movl flags,%eax
++ popl %ebx
+
+ cpu_flags.L1:
+ ret
+--- video/mmxidct_asm.S
++++ video/mmxidct_asm.S
+@@ -31,11 +31,6 @@ preSC: .short 16384,22725,21407,19266,
+ x0005000200010001:
+ .long 0x00010001,0x00050002
+ .align 8
+- .type x0040000000000000, at object
+- .size x0040000000000000,8
+-x0040000000000000:
+- .long 0, 0x00400000
+- .align 8
+ .type x5a825a825a825a82, at object
+ .size x5a825a825a825a82,8
+ x5a825a825a825a82:
+@@ -80,8 +75,21 @@ scratch7:
+ x0:
+ .long 0,0
+ .align 8
++
+ .text
+ .align 4
++
++#ifdef __PIC__
++# undef __i686 /* gcc define gets in our way */
++# define MUNG(sym) sym ## @GOTOFF(%ebx)
++# define INIT_PIC() \
++ call __i686.get_pc_thunk.bx ; \
++ addl $_GLOBAL_OFFSET_TABLE_, %ebx
++#else
++# define MUNG(sym) sym
++# define INIT_PIC()
++#endif
++
+ .globl IDCT_mmx
+ .type IDCT_mmx, at function
+ IDCT_mmx:
+@@ -92,8 +100,9 @@ IDCT_mmx:
+ pushl %edx
+ pushl %esi
+ pushl %edi
++ INIT_PIC()
+ movl 8(%ebp),%esi /* source matrix */
+- leal preSC, %ecx
++ leal MUNG(preSC), %ecx
+ /* column 0: even part
+ * use V4, V12, V0, V8 to produce V22..V25
+ */
+@@ -109,7 +118,7 @@ IDCT_mmx:
+ movq %mm1, %mm2 /* added 11/1/96 */
+ pmulhw 8*8(%esi),%mm5 /* V8 */
+ psubsw %mm0, %mm1 /* V16 */
+- pmulhw x5a825a825a825a82, %mm1 /* 23170 ->V18 */
++ pmulhw MUNG(x5a825a825a825a82), %mm1 /* 23170 ->V18 */
+ paddsw %mm0, %mm2 /* V17 */
+ movq %mm2, %mm0 /* duplicate V17 */
+ psraw $1, %mm2 /* t75=t82 */
+@@ -150,7 +159,7 @@ IDCT_mmx:
+ paddsw %mm0, %mm3 /* V29 ; free mm0 */
+ movq %mm7, %mm1 /* duplicate V26 */
+ psraw $1, %mm3 /* t91=t94 */
+- pmulhw x539f539f539f539f,%mm7 /* V33 */
++ pmulhw MUNG(x539f539f539f539f),%mm7 /* V33 */
+ psraw $1, %mm1 /* t96 */
+ movq %mm5, %mm0 /* duplicate V2 */
+ psraw $2, %mm4 /* t85=t87 */
+@@ -158,15 +167,15 @@ IDCT_mmx:
+ psubsw %mm4, %mm0 /* V28 ; free mm4 */
+ movq %mm0, %mm2 /* duplicate V28 */
+ psraw $1, %mm5 /* t90=t93 */
+- pmulhw x4546454645464546,%mm0 /* V35 */
++ pmulhw MUNG(x4546454645464546),%mm0 /* V35 */
+ psraw $1, %mm2 /* t97 */
+ movq %mm5, %mm4 /* duplicate t90=t93 */
+ psubsw %mm2, %mm1 /* V32 ; free mm2 */
+- pmulhw x61f861f861f861f8,%mm1 /* V36 */
++ pmulhw MUNG(x61f861f861f861f8),%mm1 /* V36 */
+ psllw $1, %mm7 /* t107 */
+ paddsw %mm3, %mm5 /* V31 */
+ psubsw %mm3, %mm4 /* V30 ; free mm3 */
+- pmulhw x5a825a825a825a82,%mm4 /* V34 */
++ pmulhw MUNG(x5a825a825a825a82),%mm4 /* V34 */
+ nop
+ psubsw %mm1, %mm0 /* V38 */
+ psubsw %mm7, %mm1 /* V37 ; free mm7 */
+@@ -233,7 +242,7 @@ IDCT_mmx:
+ psubsw %mm7, %mm1 /* V50 */
+ pmulhw 8*9(%esi), %mm5 /* V9 */
+ paddsw %mm7, %mm2 /* V51 */
+- pmulhw x5a825a825a825a82, %mm1 /* 23170 ->V52 */
++ pmulhw MUNG(x5a825a825a825a82), %mm1 /* 23170 ->V52 */
+ movq %mm2, %mm6 /* duplicate V51 */
+ psraw $1, %mm2 /* t138=t144 */
+ movq %mm3, %mm4 /* duplicate V1 */
+@@ -274,11 +283,11 @@ IDCT_mmx:
+ * even more by doing the correction step in a later stage when the number
+ * is actually multiplied by 16
+ */
+- paddw x0005000200010001, %mm4
++ paddw MUNG(x0005000200010001), %mm4
+ psubsw %mm6, %mm3 /* V60 ; free mm6 */
+ psraw $1, %mm0 /* t154=t156 */
+ movq %mm3, %mm1 /* duplicate V60 */
+- pmulhw x539f539f539f539f, %mm1 /* V67 */
++ pmulhw MUNG(x539f539f539f539f), %mm1 /* V67 */
+ movq %mm5, %mm6 /* duplicate V3 */
+ psraw $2, %mm4 /* t148=t150 */
+ paddsw %mm4, %mm5 /* V61 */
+@@ -287,13 +296,13 @@ IDCT_mmx:
+ psllw $1, %mm1 /* t169 */
+ paddsw %mm0, %mm5 /* V65 -> result */
+ psubsw %mm0, %mm4 /* V64 ; free mm0 */
+- pmulhw x5a825a825a825a82, %mm4 /* V68 */
++ pmulhw MUNG(x5a825a825a825a82), %mm4 /* V68 */
+ psraw $1, %mm3 /* t158 */
+ psubsw %mm6, %mm3 /* V66 */
+ movq %mm5, %mm2 /* duplicate V65 */
+- pmulhw x61f861f861f861f8, %mm3 /* V70 */
++ pmulhw MUNG(x61f861f861f861f8), %mm3 /* V70 */
+ psllw $1, %mm6 /* t165 */
+- pmulhw x4546454645464546, %mm6 /* V69 */
++ pmulhw MUNG(x4546454645464546), %mm6 /* V69 */
+ psraw $1, %mm2 /* t172 */
+ /* moved from next block */
+ movq 8*5(%esi), %mm0 /* V56 */
+@@ -418,7 +427,7 @@ IDCT_mmx:
+ * movq 8*13(%esi), %mm4 tmt13
+ */
+ psubsw %mm4, %mm3 /* V134 */
+- pmulhw x5a825a825a825a82, %mm3 /* 23170 ->V136 */
++ pmulhw MUNG(x5a825a825a825a82), %mm3 /* 23170 ->V136 */
+ movq 8*9(%esi), %mm6 /* tmt9 */
+ paddsw %mm4, %mm5 /* V135 ; mm4 free */
+ movq %mm0, %mm4 /* duplicate tmt1 */
+@@ -447,17 +456,17 @@ IDCT_mmx:
+ psubsw %mm7, %mm0 /* V144 */
+ movq %mm0, %mm3 /* duplicate V144 */
+ paddsw %mm7, %mm2 /* V147 ; free mm7 */
+- pmulhw x539f539f539f539f, %mm0 /* 21407-> V151 */
++ pmulhw MUNG(x539f539f539f539f), %mm0 /* 21407-> V151 */
+ movq %mm1, %mm7 /* duplicate tmt3 */
+ paddsw %mm5, %mm7 /* V145 */
+ psubsw %mm5, %mm1 /* V146 ; free mm5 */
+ psubsw %mm1, %mm3 /* V150 */
+ movq %mm7, %mm5 /* duplicate V145 */
+- pmulhw x4546454645464546, %mm1 /* 17734-> V153 */
++ pmulhw MUNG(x4546454645464546), %mm1 /* 17734-> V153 */
+ psubsw %mm2, %mm5 /* V148 */
+- pmulhw x61f861f861f861f8, %mm3 /* 25080-> V154 */
++ pmulhw MUNG(x61f861f861f861f8), %mm3 /* 25080-> V154 */
+ psllw $2, %mm0 /* t311 */
+- pmulhw x5a825a825a825a82, %mm5 /* 23170-> V152 */
++ pmulhw MUNG(x5a825a825a825a82), %mm5 /* 23170-> V152 */
+ paddsw %mm2, %mm7 /* V149 ; free mm2 */
+ psllw $1, %mm1 /* t313 */
+ nop /* without the nop - freeze here for one clock */
+@@ -483,7 +492,7 @@ IDCT_mmx:
+ paddsw %mm3, %mm6 /* V164 ; free mm3 */
+ movq %mm4, %mm3 /* duplicate V142 */
+ psubsw %mm5, %mm4 /* V165 ; free mm5 */
+- movq %mm2, scratch7 /* out7 */
++ movq %mm2, MUNG(scratch7) /* out7 */
+ psraw $4, %mm6
+ psraw $4, %mm4
+ paddsw %mm5, %mm3 /* V162 */
+@@ -494,11 +503,11 @@ IDCT_mmx:
+ */
+ movq %mm6, 8*9(%esi) /* out9 */
+ paddsw %mm1, %mm0 /* V161 */
+- movq %mm3, scratch5 /* out5 */
++ movq %mm3, MUNG(scratch5) /* out5 */
+ psubsw %mm1, %mm5 /* V166 ; free mm1 */
+ movq %mm4, 8*11(%esi) /* out11 */
+ psraw $4, %mm5
+- movq %mm0, scratch3 /* out3 */
++ movq %mm0, MUNG(scratch3) /* out3 */
+ movq %mm2, %mm4 /* duplicate V140 */
+ movq %mm5, 8*13(%esi) /* out13 */
+ paddsw %mm7, %mm2 /* V160 */
+@@ -508,7 +517,7 @@ IDCT_mmx:
+ /* moved from the next block */
+ movq 8*3(%esi), %mm7
+ psraw $4, %mm4
+- movq %mm2, scratch1 /* out1 */
++ movq %mm2, MUNG(scratch1) /* out1 */
+ /* moved from the next block */
+ movq %mm0, %mm1
+ movq %mm4, 8*15(%esi) /* out15 */
+@@ -565,15 +574,15 @@ IDCT_mmx:
+ paddsw %mm4, %mm3 /* V113 ; free mm4 */
+ movq %mm0, %mm4 /* duplicate V110 */
+ paddsw %mm1, %mm2 /* V111 */
+- pmulhw x539f539f539f539f, %mm0 /* 21407-> V117 */
++ pmulhw MUNG(x539f539f539f539f), %mm0 /* 21407-> V117 */
+ psubsw %mm1, %mm5 /* V112 ; free mm1 */
+ psubsw %mm5, %mm4 /* V116 */
+ movq %mm2, %mm1 /* duplicate V111 */
+- pmulhw x4546454645464546, %mm5 /* 17734-> V119 */
++ pmulhw MUNG(x4546454645464546), %mm5 /* 17734-> V119 */
+ psubsw %mm3, %mm2 /* V114 */
+- pmulhw x61f861f861f861f8, %mm4 /* 25080-> V120 */
++ pmulhw MUNG(x61f861f861f861f8), %mm4 /* 25080-> V120 */
+ paddsw %mm3, %mm1 /* V115 ; free mm3 */
+- pmulhw x5a825a825a825a82, %mm2 /* 23170-> V118 */
++ pmulhw MUNG(x5a825a825a825a82), %mm2 /* 23170-> V118 */
+ psllw $2, %mm0 /* t266 */
+ movq %mm1, (%esi) /* save V115 */
+ psllw $1, %mm5 /* t268 */
+@@ -591,7 +600,7 @@ IDCT_mmx:
+ movq %mm6, %mm3 /* duplicate tmt4 */
+ psubsw %mm0, %mm6 /* V100 */
+ paddsw %mm0, %mm3 /* V101 ; free mm0 */
+- pmulhw x5a825a825a825a82, %mm6 /* 23170 ->V102 */
++ pmulhw MUNG(x5a825a825a825a82), %mm6 /* 23170 ->V102 */
+ movq %mm7, %mm5 /* duplicate tmt0 */
+ movq 8*8(%esi), %mm1 /* tmt8 */
+ paddsw %mm1, %mm7 /* V103 */
+@@ -625,10 +634,10 @@ IDCT_mmx:
+ movq 8*2(%esi), %mm3 /* V123 */
+ paddsw %mm4, %mm7 /* out0 */
+ /* moved up from next block */
+- movq scratch3, %mm0
++ movq MUNG(scratch3), %mm0
+ psraw $4, %mm7
+ /* moved up from next block */
+- movq scratch5, %mm6
++ movq MUNG(scratch5), %mm6
+ psubsw %mm4, %mm1 /* out14 ; free mm4 */
+ paddsw %mm3, %mm5 /* out2 */
+ psraw $4, %mm1
+@@ -639,7 +648,7 @@ IDCT_mmx:
+ movq %mm5, 8*2(%esi) /* out2 ; free mm5 */
+ psraw $4, %mm2
+ /* moved up to the prev block */
+- movq scratch7, %mm4
++ movq MUNG(scratch7), %mm4
+ /* moved up to the prev block */
+ psraw $4, %mm0
+ movq %mm2, 8*12(%esi) /* out12 ; free mm2 */
+@@ -647,13 +656,13 @@ IDCT_mmx:
+ psraw $4, %mm6
+ /* move back the data to its correct place
+ * moved up to the prev block
+- * movq scratch3, %mm0
+- * movq scratch5, %mm6
+- * movq scratch7, %mm4
++ * movq MUNG(scratch3), %mm0
++ * movq MUNG(scratch5), %mm6
++ * movq MUNG(scratch7), %mm4
+ * psraw $4, %mm0
+ * psraw $4, %mm6
+ */
+- movq scratch1, %mm1
++ movq MUNG(scratch1), %mm1
+ psraw $4, %mm4
+ movq %mm0, 8*3(%esi) /* out3 */
+ psraw $4, %mm1
+@@ -671,6 +680,15 @@ IDCT_mmx:
+ .Lfe1:
+ .size IDCT_mmx,.Lfe1-IDCT_mmx
+
++#ifdef __PIC__
++ .section .gnu.linkonce.t.__i686.get_pc_thunk.bx,"ax", at progbits
++.globl __i686.get_pc_thunk.bx
++ .hidden __i686.get_pc_thunk.bx
++ .type __i686.get_pc_thunk.bx, at function
++ __i686.get_pc_thunk.bx:
++ movl (%esp), %ebx
++ ret
++#endif
+
+ #endif /* i386 && USE_MMX */
+
================================================================
More information about the pld-cvs-commit
mailing list