X-Git-Url: http://git.megacz.com/?a=blobdiff_plain;f=ghc%2Fdriver%2Fmangler%2Fghc-asm.lprl;h=760175a5751c22f47fe5e305db24f1ccec8af787;hb=6f917d9600e7e5ba63a33f4079bcbc6ebbdb10fb;hp=dfac27620f49d711b606fe77fa0ee3e239589115;hpb=1da969c52be2e56893625957162cf56dd1168ec9;p=ghc-hetmet.git diff --git a/ghc/driver/mangler/ghc-asm.lprl b/ghc/driver/mangler/ghc-asm.lprl index dfac276..760175a 100644 --- a/ghc/driver/mangler/ghc-asm.lprl +++ b/ghc/driver/mangler/ghc-asm.lprl @@ -99,20 +99,16 @@ sub init_TARGET_STUFF { $T_MOVE_DIRVS = '^(\s*(\$.*\.\.ng:|\.align\s+\d+|\.(globl|ent)\s+\S+|\#.*|\.(file|loc)\s+\S+\s+\S+|\.text|\.r?data)\n)'; $T_COPY_DIRVS = '^\s*(\$.*\.\.ng:|\#|\.(file|globl|ent|loc))'; - $T_hsc_cc_PAT = '\.ascii.*\)(hsc|cc) (.*)\\\\11"\n\t\.ascii\s+"(.*)\\\\0"'; $T_DOT_WORD = '\.(long|quad|byte|word)'; $T_DOT_GLOBAL = '^\t\.globl'; $T_HDR_literal = "\.rdata\n\t\.align 3\n"; $T_HDR_misc = "\.text\n\t\.align 3\n"; $T_HDR_data = "\.data\n\t\.align 3\n"; - $T_HDR_consist = "\.text\n"; + $T_HDR_rodata = "\.rdata\n\t\.align 3\n"; $T_HDR_closure = "\.data\n\t\.align 3\n"; - $T_HDR_srt = "\.text\n\t\.align 3\n"; $T_HDR_info = "\.text\n\t\.align 3\n"; $T_HDR_entry = "\.text\n\t\.align 3\n"; $T_HDR_vector = "\.text\n\t\.align 3\n"; - $T_HDR_direct = "\.text\n\t\.align 3\n"; - $T_create_word = "\t.quad"; #--------------------------------------------------------# } elsif ( $TargetPlatform =~ /^hppa/ ) { @@ -126,23 +122,19 @@ sub init_TARGET_STUFF { $T_MOVE_DIRVS = '^((\s+\.(IMPORT|EXPORT|PARAM).*|\s+\.align\s+\d+|\s+\.(SPACE|SUBSPA)\s+\S+|\s*)\n)'; $T_COPY_DIRVS = '^\s+\.(IMPORT|EXPORT)'; - $T_hsc_cc_PAT = '\.STRING.*\)(hsc|cc) (.*)\\\\x09(.*)\\\\x00"'; $T_DOT_WORD = '\.(blockz|word|half|byte)'; $T_DOT_GLOBAL = '^\s+\.EXPORT'; $T_HDR_literal = "\t.SPACE \$TEXT\$\n\t.SUBSPA \$LIT\$\n"; $T_HDR_misc = "\t.SPACE \$TEXT\$\n\t.SUBSPA \$CODE\$\n\t\.align 4\n"; $T_HDR_data = "\t.SPACE \$PRIVATE\$\n\t.SUBSPA \$DATA\$\n\t\.align 4\n"; - $T_HDR_consist = "\t.SPACE \$TEXT\$\n\t.SUBSPA \$LIT\$\n"; + $T_HDR_rodata = "\t.SPACE \$PRIVATE\$\n\t.SUBSPA \$DATA\$\n\t\.align 4\n"; $T_HDR_closure = "\t.SPACE \$PRIVATE\$\n\t.SUBSPA \$DATA\$\n\t\.align 4\n"; - $T_HDR_srt = "\t.SPACE \$PRIVATE\$\n\t.SUBSPA \$DATA\$\n\t\.align 4\n"; $T_HDR_info = "\t.SPACE \$TEXT\$\n\t.SUBSPA \$CODE\$\n\t\.align 4\n"; $T_HDR_entry = "\t.SPACE \$TEXT\$\n\t.SUBSPA \$CODE\$\n\t\.align 4\n"; $T_HDR_vector = "\t.SPACE \$TEXT\$\n\t.SUBSPA \$CODE\$\n\t\.align 4\n"; - $T_HDR_direct = "\t.SPACE \$TEXT\$\n\t.SUBSPA \$CODE\$\n\t\.align 4\n"; - $T_create_word = "\t.word"; #--------------------------------------------------------# - } elsif ( $TargetPlatform =~ /^i386-.*-(linuxaout|freebsd2|openbsd|nextstep3|cygwin32|mingw32)$/ ) { + } elsif ( $TargetPlatform =~ /^i386-.*-(linuxaout|freebsd2|nextstep3|cygwin32|mingw32)$/ ) { # NeXT added but not tested. CaS $T_STABBY = 1; # 1 iff .stab things (usually if a.out format) @@ -154,58 +146,45 @@ sub init_TARGET_STUFF { $T_X86_PRE_LLBL = 'L'; $T_X86_BADJMP = '^\tjmp [^L\*]'; - $T_MOVE_DIRVS = '^(\s*(\.(p2)?align\s+\d+(,0x90)?|\.globl\s+\S+|\.text|\.data|\.stab[^n].*|\.type\s+.*|\.size\s+.*)\n)'; - $T_COPY_DIRVS = '\.(globl|stab)'; - $T_hsc_cc_PAT = '\.ascii.*\)(hsc|cc) (.*)\\\\11"\n\t\.ascii\s+"(.*)\\\\0"'; + $T_MOVE_DIRVS = '^(\s*(\.(p2)?align\s.*|\.globl\s+\S+|\.text|\.data|\.stab[^n].*|\.type\s+.*|\.size\s+.*|\.lcomm.*)\n)'; + $T_COPY_DIRVS = '\.(globl|stab|lcomm)'; $T_DOT_WORD = '\.(long|word|value|byte|space)'; $T_DOT_GLOBAL = '\.globl'; $T_HDR_literal = "\.text\n\t\.align 2\n"; $T_HDR_misc = "\.text\n\t\.align 2,0x90\n"; $T_HDR_data = "\.data\n\t\.align 2\n"; - $T_HDR_consist = "\.text\n"; + $T_HDR_rodata = "\.text\n\t\.align 2\n"; $T_HDR_closure = "\.data\n\t\.align 2\n"; - $T_HDR_srt = "\.text\n\t\.align 2\n"; $T_HDR_info = "\.text\n\t\.align 2\n"; # NB: requires padding $T_HDR_entry = "\.text\n"; # no .align so we're right next to _info (arguably wrong...?) $T_HDR_vector = "\.text\n\t\.align 2\n"; # NB: requires padding - $T_HDR_direct = "\.text\n\t\.align 2,0x90\n"; - $T_create_word = "\t.word"; #--------------------------------------------------------# - } elsif ( $TargetPlatform =~ /^i386-.*-(solaris2|linux|freebsd|netbsd)$/ ) { + } elsif ( $TargetPlatform =~ /^i386-.*-(solaris2|linux|gnu|freebsd|netbsd|openbsd|kfreebsdgnu)$/ ) { $T_STABBY = 0; # 1 iff .stab things (usually if a.out format) $T_US = ''; # _ if symbols have an underscore on the front $T_PRE_APP = # regexp that says what comes before APP/NO_APP - ($TargetPlatform =~ /-(linux|freebsd|netbsd)$/) ? '#' : '/' ; + ($TargetPlatform =~ /-(linux|gnu|freebsd|netbsd|openbsd)$/) ? '#' : '/' ; $T_CONST_LBL = '^\.LC(\d+):$'; # regexp for what such a lbl looks like $T_POST_LBL = ':'; $T_X86_PRE_LLBL_PAT = '\.L'; $T_X86_PRE_LLBL = '.L'; $T_X86_BADJMP = '^\tjmp\s+[^\.\*]'; - $T_MOVE_DIRVS = '^(\s*(\.(p2)?align\s+\d+(,\s*0x90)?|\.globl\s+\S+|\.text|\.data|\.section\s+.*|\.type\s+.*|\.size\s+\S+,\d+|\.ident.*)\n)'; - $T_COPY_DIRVS = '^\s*\.(globl|type|size)'; - - if ( $TargetPlatform =~ /freebsd|netbsd/ ) { - $T_hsc_cc_PAT = '\.ascii.*\)(hsc|cc) (.*)\\\\11"\n\t\.ascii\s+"(.*)\\\\0"'; - } else { - $T_hsc_cc_PAT = '\.string.*\)(hsc|cc) (.*)\\\\t(.*)"'; - } + $T_MOVE_DIRVS = '^(\s*(\.(p2)?align\s.*|\.globl\s+\S+|\.text|\.data|\.section\s+.*|\.type\s+.*|\.size\s+\S+\s*,\s*\d+|\.ident.*|\.local.*)\n)'; + $T_COPY_DIRVS = '^\s*\.(globl|type|size|local)'; $T_DOT_WORD = '\.(long|value|word|byte|zero)'; $T_DOT_GLOBAL = '\.globl'; $T_HDR_literal = "\.section\t\.rodata\n"; # or just use .text??? (WDP 95/11) $T_HDR_misc = "\.text\n\t\.align 4\n"; - $T_HDR_data = "\.data\n\t\.align 4\n"; # ToDo: change align?? - $T_HDR_consist = "\.text\n"; - $T_HDR_closure = "\.data\n\t\.align 4\n"; # ToDo: change align? - $T_HDR_srt = "\.text\n\t\.align 4\n"; # ToDo: change align? - $T_HDR_info = "\.text\n\t\.align 4\n"; # NB: requires padding + $T_HDR_data = "\.data\n\t\.align 4\n"; + $T_HDR_rodata = "\.section\t\.rodata\n\t\.align 4\n"; + $T_HDR_closure = "\.data\n\t\.align 4\n"; + $T_HDR_info = "\.text\n\t\.align 4\n"; $T_HDR_entry = "\.text\n"; # no .align so we're right next to _info (arguably wrong...?) $T_HDR_vector = "\.text\n\t\.align 4\n"; # NB: requires padding - $T_HDR_direct = "\.text\n\t\.align 4\n"; - $T_create_word = "\t.word"; #--------------------------------------------------------# } elsif ( $TargetPlatform =~ /^ia64-.*-linux$/ ) { @@ -219,20 +198,50 @@ sub init_TARGET_STUFF { $T_MOVE_DIRVS = '^(\s*\.(global|proc|pred\.safe_across_calls|text|data|section|subsection|align|size|type|ident)\s+.*\n)'; $T_COPY_DIRVS = '\.(global|proc)'; - $T_hsc_cc_PAT = '\.string.*\)(hsc|cc) (.*)\\\\t(.*)"'; $T_DOT_WORD = '\.(long|value|byte|zero)'; $T_DOT_GLOBAL = '\.global'; $T_HDR_literal = "\.section\t\.rodata\n"; $T_HDR_misc = "\.text\n\t\.align 8\n"; $T_HDR_data = "\.data\n\t\.align 8\n"; - $T_HDR_consist = "\.text\n"; + $T_HDR_rodata = "\.section\t\.rodata\n\t\.align 8\n"; $T_HDR_closure = "\.data\n\t\.align 8\n"; - $T_HDR_srt = "\.text\n\t\.align 8\n"; $T_HDR_info = "\.text\n\t\.align 8\n"; $T_HDR_entry = "\.text\n\t\.align 16\n"; $T_HDR_vector = "\.text\n\t\.align 8\n"; - $T_HDR_direct = "\.text\n\t\.align 8\n"; - $T_create_word = "\t.word"; + + #--------------------------------------------------------# + } elsif ( $TargetPlatform =~ /^x86_64-.*-(linux|openbsd)$/ ) { + + $T_STABBY = 0; # 1 iff .stab things (usually if a.out format) + $T_US = ''; # _ if symbols have an underscore on the front + $T_PRE_APP = '#'; + $T_CONST_LBL = '^\.LC(\d+):$'; # regexp for what such a lbl looks like + $T_POST_LBL = ':'; + + $T_MOVE_DIRVS = '^(\s*\.(globl|text|data|section|align|size|type|ident|local)\s+.*\n)'; + $T_COPY_DIRVS = '\.(globl|local)'; + + $T_DOT_WORD = '\.(quad|long|value|byte|zero)'; + $T_DOT_GLOBAL = '\.global'; + + $T_HDR_literal16 = "\.section\t\.rodata.cst16\n\t.align 16\n"; + $T_HDR_literal = "\.section\t\.rodata\n"; + + $T_HDR_misc = "\.text\n\t\.align 8\n"; + $T_HDR_data = "\.data\n\t\.align 8\n"; + $T_HDR_rodata = "\.section\t\.rodata\n\t\.align 8\n"; + + # the assembler on x86_64/Linux refuses to generate code for + # .quad x - y + # where x is in the text section and y in the rodata section. + # It works if y is in the text section, though. This is probably + # going to cause difficulties for PIC, I imagine. + $T_HDR_relrodata= "\.text\n\t\.align 8\n"; + + $T_HDR_closure = "\.data\n\t\.align 8\n"; + $T_HDR_info = "\.text\n\t\.align 8\n"; + $T_HDR_entry = "\.text\n\t\.align 8\n"; + $T_HDR_vector = "\.text\n\t\.align 8\n"; #--------------------------------------------------------# } elsif ( $TargetPlatform =~ /^m68k-.*-sunos4/ ) { @@ -245,21 +254,17 @@ sub init_TARGET_STUFF { $T_MOVE_DIRVS = '^(\s*(\.align\s+\d+|\.proc\s+\d+|\.const|\.cstring|\.globl\s+\S+|\.text|\.data|\.even|\.stab[^n].*)\n)'; $T_COPY_DIRVS = '\.(globl|proc|stab)'; - $T_hsc_cc_PAT = '\.ascii.*\)(hsc|cc) (.*)\\\\11"\n\t\.ascii\s+"(.*)\\\\0"'; $T_DOT_WORD = '\.long'; $T_DOT_GLOBAL = '\.globl'; $T_HDR_literal = "\.text\n\t\.even\n"; $T_HDR_misc = "\.text\n\t\.even\n"; $T_HDR_data = "\.data\n\t\.even\n"; - $T_HDR_consist = "\.text\n"; + $T_HDR_rodata = "\.text\n\t\.even\n"; $T_HDR_closure = "\.data\n\t\.even\n"; - $T_HDR_srt = "\.text\n\t\.even\n"; $T_HDR_info = "\.text\n\t\.even\n"; $T_HDR_entry = "\.text\n\t\.even\n"; $T_HDR_vector = "\.text\n\t\.even\n"; - $T_HDR_direct = "\.text\n\t\.even\n"; - $T_create_word = "\t.long"; #--------------------------------------------------------# } elsif ( $TargetPlatform =~ /^mips-.*/ ) { @@ -273,51 +278,121 @@ sub init_TARGET_STUFF { $T_MOVE_DIRVS = '^(\s*(\.align\s+\d+|\.(globl|ent)\s+\S+|\.text|\.r?data)\n)'; $T_COPY_DIRVS = '\.(globl|ent)'; - $T_hsc_cc_PAT = 'I WAS TOO LAZY TO DO THIS BIT (WDP 95/05)'; $T_DOT_WORD = '\.word'; $T_DOT_GLOBAL = '^\t\.globl'; $T_HDR_literal = "\t\.rdata\n\t\.align 2\n"; $T_HDR_misc = "\t\.text\n\t\.align 2\n"; $T_HDR_data = "\t\.data\n\t\.align 2\n"; - $T_HDR_consist = 'TOO LAZY TO DO THIS TOO'; + $T_HDR_rodata = "\t\.rdata\n\t\.align 2\n"; $T_HDR_closure = "\t\.data\n\t\.align 2\n"; - $T_HDR_srt = "\t\.text\n\t\.align 2\n"; $T_HDR_info = "\t\.text\n\t\.align 2\n"; $T_HDR_entry = "\t\.text\n\t\.align 2\n"; $T_HDR_vector = "\t\.text\n\t\.align 2\n"; - $T_HDR_direct = "\t\.text\n\t\.align 2\n"; - $T_create_word = "\t.word"; #--------------------------------------------------------# - } elsif ( $TargetPlatform =~ /^powerpc-apple-.*/ ) { + } elsif ( $TargetPlatform =~ /^powerpc-apple-darwin.*/ ) { # Apple PowerPC Darwin/MacOS X. $T_STABBY = 0; # 1 iff .stab things (usually if a.out format) $T_US = '_'; # _ if symbols have an underscore on the front - $T_PRE_APP = 'WHAT IS THIS'; # regexp that says what comes before APP/NO_APP + $T_PRE_APP = 'DOESNT APPLY'; # regexp that says what comes before APP/NO_APP $T_CONST_LBL = '^\LC\d+:'; # regexp for what such a lbl looks like $T_POST_LBL = ':'; - $T_MOVE_DIRVS = '^(\s*(\.align \d+|\.text|\.data|\.const_data|\.cstring|\.non_lazy_symbol_pointer|\.const|\.static_const|\.literal4|\.literal8|\.static_data|\.globl \S+)\n)'; - $T_COPY_DIRVS = '\.(globl)'; + $T_MOVE_DIRVS = '^(\s*(\.align \d+|\.text|\.data|\.const_data|\.cstring|\.non_lazy_symbol_pointer|\.const|\.static_const|\.literal4|\.literal8|\.static_data|\.globl \S+|\.section .*|\.lcomm.*)\n)'; + $T_COPY_DIRVS = '\.(globl|lcomm)'; - $T_hsc_cc_PAT = '\.byte.*\)(hsc|cc) (.*)"\n\t\.byte \d+\n\t\.byte "(.*)"\n\t\.byte \d+'; $T_DOT_WORD = '\.(long|short|byte|fill|space)'; $T_DOT_GLOBAL = '\.globl'; $T_HDR_toc = "\.toc\n"; - $T_HDR_literal = "\t\.const_data\n\t\.align 2\n"; + $T_HDR_literal = "\t\.const\n\t\.align 2\n"; $T_HDR_misc = "\t\.text\n\t\.align 2\n"; $T_HDR_data = "\t\.data\n\t\.align 2\n"; - $T_HDR_consist = "\t\.text\n\t\.align 2\n"; - $T_HDR_closure = "\t\.const_data\n\t\.align 2\n"; - $T_HDR_srt = "\t\.text\n\t\.align 2\n"; + $T_HDR_rodata = "\t\.const\n\t\.align 2\n"; + $T_HDR_relrodata= "\t\.const_data\n\t\.align 2\n"; + $T_HDR_closure = "\t\.data\n\t\.align 2\n"; $T_HDR_info = "\t\.text\n\t\.align 2\n"; $T_HDR_entry = "\t\.text\n\t\.align 2\n"; $T_HDR_vector = "\t\.text\n\t\.align 2\n"; - $T_HDR_direct = "\t\.text\n\t\.align 2\n"; - $T_create_word = "\t.long"; #--------------------------------------------------------# - } elsif ( $TargetPlatform =~ /^sparc-.*-solaris2/ ) { + } elsif ( $TargetPlatform =~ /^i386-apple-darwin.*/ ) { + # Apple PowerPC Darwin/MacOS X. + $T_STABBY = 0; # 1 iff .stab things (usually if a.out format) + $T_US = '_'; # _ if symbols have an underscore on the front + $T_PRE_APP = 'DOESNT APPLY'; # regexp that says what comes before APP/NO_APP + $T_CONST_LBL = '^\LC\d+:'; # regexp for what such a lbl looks like + $T_POST_LBL = ':'; + $T_X86_PRE_LLBL_PAT = 'L'; + $T_X86_PRE_LLBL = 'L'; + $T_X86_BADJMP = '^\tjmp [^L\*]'; + + $T_MOVE_DIRVS = '^(\s*(\.align \d+|\.text|\.data|\.const_data|\.cstring|\.non_lazy_symbol_pointer|\.const|\.static_const|\.literal4|\.literal8|\.static_data|\.globl \S+|\.section .*|\.lcomm.*)\n)'; + $T_COPY_DIRVS = '\.(globl|lcomm)'; + + $T_DOT_WORD = '\.(long|short|byte|fill|space)'; + $T_DOT_GLOBAL = '\.globl'; + $T_HDR_toc = "\.toc\n"; + $T_HDR_literal16= "\t\.literal8\n\t\.align 4\n"; + $T_HDR_literal = "\t\.const\n\t\.align 4\n"; + $T_HDR_misc = "\t\.text\n\t\.align 2\n"; + $T_HDR_data = "\t\.data\n\t\.align 2\n"; + $T_HDR_rodata = "\t\.const\n\t\.align 2\n"; + $T_HDR_relrodata= "\t\.const_data\n\t\.align 2\n"; + $T_HDR_closure = "\t\.data\n\t\.align 2\n"; + $T_HDR_info = "\t\.text\n\t\.align 2\n"; + $T_HDR_entry = "\t\.text\n\t\.align 2\n"; + $T_HDR_vector = "\t\.text\n\t\.align 2\n"; + + #--------------------------------------------------------# + } elsif ( $TargetPlatform =~ /^powerpc-.*-linux/ ) { + # PowerPC Linux + $T_STABBY = 0; # 1 iff .stab things (usually if a.out format) + $T_US = ''; # _ if symbols have an underscore on the front + $T_PRE_APP = '^#'; # regexp that says what comes before APP/NO_APP + $T_CONST_LBL = '^\.LC\d+:'; # regexp for what such a lbl looks like + $T_POST_LBL = ':'; + + $T_MOVE_DIRVS = '^(\s*(\.(p2)?align\s+\d+(,\s*0x90)?|\.globl\s+\S+|\.text|\.data|\.section\s+.*|\.type\s+.*|\.size\s+\S+\s*,\s*\d+|\.ident.*|\.local.*)\n)'; + $T_COPY_DIRVS = '^\s*\.(globl|type|size|local)'; + + $T_DOT_WORD = '\.(long|short|byte|fill|space)'; + $T_DOT_GLOBAL = '\.globl'; + $T_HDR_toc = "\.toc\n"; + $T_HDR_literal = "\t\.section\t.rodata\n\t\.align 2\n"; + $T_HDR_misc = "\t\.text\n\t\.align 2\n"; + $T_HDR_data = "\t\.data\n\t\.align 2\n"; + $T_HDR_rodata = "\t\.section\t.rodata\n\t\.align 2\n"; + $T_HDR_closure = "\t\.data\n\t\.align 2\n"; + $T_HDR_info = "\t\.text\n\t\.align 2\n"; + $T_HDR_entry = "\t\.text\n\t\.align 2\n"; + $T_HDR_vector = "\t\.text\n\t\.align 2\n"; + + #--------------------------------------------------------# + } elsif ( $TargetPlatform =~ /^powerpc64-.*-linux/ ) { + # PowerPC 64 Linux + $T_STABBY = 0; # 1 iff .stab things (usually if a.out format) + $T_US = '\.'; # _ if symbols have an underscore on the front + $T_PRE_APP = '^#'; # regexp that says what comes before APP/NO_APP + $T_CONST_LBL = '^\.LC\d+:'; # regexp for what such a lbl looks like + $T_POST_LBL = ':'; + + $T_MOVE_DIRVS = '^(\s*(\.(p2)?align\s+\d+(,\s*0x90)?|\.globl\s+\S+|\.text|\.data|\.section\s+.*|\.type\s+.*|\.size\s+\S+\s*,\s*\d+|\.ident.*|\.local.*)\n)'; + $T_COPY_DIRVS = '^\s*\.(globl|type|size|local)'; + + $T_DOT_WORD = '\.(long|short|byte|fill|space)'; + $T_DOT_GLOBAL = '\.globl'; + $T_HDR_toc = "\.toc\n"; + $T_HDR_literal = "\t\.section\t\".toc\",\"aw\"\n"; + $T_HDR_misc = "\t\.text\n\t\.align 2\n"; + $T_HDR_data = "\t\.data\n\t\.align 2\n"; + $T_HDR_rodata = "\t\.section\t.rodata\n\t\.align 2\n"; + $T_HDR_closure = "\t\.data\n\t\.align 2\n"; + $T_HDR_info = "\t\.text\n\t\.align 2\n"; + $T_HDR_entry = "\t\.text\n\t\.align 2\n"; + $T_HDR_vector = "\t\.text\n\t\.align 2\n"; + + #--------------------------------------------------------# + } elsif ( $TargetPlatform =~ /^sparc-.*-(solaris2|openbsd)/ ) { $T_STABBY = 0; # 1 iff .stab things (usually if a.out format) $T_US = ''; # _ if symbols have an underscore on the front @@ -325,23 +400,19 @@ sub init_TARGET_STUFF { $T_CONST_LBL = '^\.LLC(\d+):$'; # regexp for what such a lbl looks like $T_POST_LBL = ':'; - $T_MOVE_DIRVS = '^((\s+\.align\s+\d+|\s+\.proc\s+\d+|\s+\.global\s+\S+|\.text|\.data|\.stab.*|\.section.*|\s+\.type.*|\s+\.size.*)\n)'; - $T_COPY_DIRVS = '\.(global|proc|stab)'; + $T_MOVE_DIRVS = '^((\s+\.align\s+\d+|\s+\.proc\s+\d+|\s+\.global\s+\S+|\s+\.local\s+\S+|\.text|\.data|\.stab.*|\s*\.section.*|\s+\.type.*|\s+\.size.*)\n)'; + $T_COPY_DIRVS = '\.(global|local|proc|stab)'; - $T_hsc_cc_PAT = '\.asciz.*\)(hsc|cc) (.*)\\\\t(.*)"'; $T_DOT_WORD = '\.(long|word|byte|half|skip|uahalf|uaword)'; $T_DOT_GLOBAL = '^\t\.global'; $T_HDR_literal = "\.text\n\t\.align 8\n"; $T_HDR_misc = "\.text\n\t\.align 4\n"; $T_HDR_data = "\.data\n\t\.align 8\n"; - $T_HDR_consist = "\.text\n"; + $T_HDR_rodata = "\.text\n\t\.align 4\n"; $T_HDR_closure = "\.data\n\t\.align 4\n"; - $T_HDR_srt = "\.data\n\t\.align 4\n"; - $T_HDR_info = "\.text\n\t\.align 4\n"; + $T_HDR_info = "\.text\n\t\.align 4\n"; $T_HDR_entry = "\.text\n\t\.align 4\n"; $T_HDR_vector = "\.text\n\t\.align 4\n"; - $T_HDR_direct = "\.text\n\t\.align 4\n"; - $T_create_word = "\t.word"; #--------------------------------------------------------# } elsif ( $TargetPlatform =~ /^sparc-.*-sunos4/ ) { @@ -354,21 +425,40 @@ sub init_TARGET_STUFF { $T_MOVE_DIRVS = '^((\s+\.align\s+\d+|\s+\.proc\s+\d+|\s+\.global\s+\S+|\.text|\.data|\.stab.*)\n)'; $T_COPY_DIRVS = '\.(global|proc|stab)'; - $T_hsc_cc_PAT = '\.ascii.*\)(hsc|cc) (.*)\\\\11"\n\t\.ascii\s+"(.*)\\\\0"'; $T_DOT_WORD = '\.word'; $T_DOT_GLOBAL = '^\t\.global'; $T_HDR_literal = "\.text\n\t\.align 8\n"; $T_HDR_misc = "\.text\n\t\.align 4\n"; $T_HDR_data = "\.data\n\t\.align 8\n"; - $T_HDR_consist = "\.text\n"; + $T_HDR_rodata = "\.text\n\t\.align 4\n"; $T_HDR_closure = "\.data\n\t\.align 4\n"; - $T_HDR_srt = "\.data\n\t\.align 4\n"; $T_HDR_info = "\.text\n\t\.align 4\n"; $T_HDR_entry = "\.text\n\t\.align 4\n"; $T_HDR_vector = "\.text\n\t\.align 4\n"; - $T_HDR_direct = "\.text\n\t\.align 4\n"; - $T_create_word = "\t.word"; + + #--------------------------------------------------------# + } elsif ( $TargetPlatform =~ /^sparc-.*-linux/ ) { + $T_STABBY = 0; # 1 iff .stab things (usually if a.out format) + $T_US = ''; # _ if symbols have an underscore on the front + $T_PRE_APP = '#'; # regexp that says what comes before APP/NO_APP + # Probably doesn't apply anyway + $T_CONST_LBL = '^\.LLC(\d+):$'; # regexp for what such a lbl looks like + $T_POST_LBL = ':'; + + $T_MOVE_DIRVS = '^((\s+\.align\s+\d+|\s+\.proc\s+\d+|\s+\.global\s+\S+|\.text|\.data|\.seg|\.stab.*|\s+?\.section.*|\s+\.type.*|\s+\.size.*)\n)'; + $T_COPY_DIRVS = '\.(global|globl|proc|stab)'; + + $T_DOT_WORD = '\.(long|word|nword|xword|byte|half|short|skip|uahalf|uaword)'; + $T_DOT_GLOBAL = '^\t\.global'; + $T_HDR_literal = "\.text\n\t\.align 8\n"; + $T_HDR_misc = "\.text\n\t\.align 4\n"; + $T_HDR_data = "\.data\n\t\.align 8\n"; + $T_HDR_rodata = "\.text\n\t\.align 4\n"; + $T_HDR_closure = "\.data\n\t\.align 4\n"; + $T_HDR_info = "\.text\n\t\.align 4\n"; + $T_HDR_entry = "\.text\n\t\.align 4\n"; + $T_HDR_vector = "\.text\n\t\.align 4\n"; #--------------------------------------------------------# } else { @@ -376,6 +466,12 @@ sub init_TARGET_STUFF { exit 1; } + if($T_HDR_relrodata eq "") { + # default values: + # relrodata defaults to rodata. + $T_HDR_relrodata = $T_HDR_rodata; + } + if ( 0 ) { print STDERR "T_STABBY: $T_STABBY\n"; print STDERR "T_US: $T_US\n"; @@ -389,17 +485,15 @@ if ( $TargetPlatform =~ /^i386-/ ) { } print STDERR "T_MOVE_DIRVS: $T_MOVE_DIRVS\n"; print STDERR "T_COPY_DIRVS: $T_COPY_DIRVS\n"; -print STDERR "T_hsc_cc_PAT: $T_hsc_cc_PAT\n"; print STDERR "T_DOT_WORD: $T_DOT_WORD\n"; print STDERR "T_HDR_literal: $T_HDR_literal\n"; print STDERR "T_HDR_misc: $T_HDR_misc\n"; print STDERR "T_HDR_data: $T_HDR_data\n"; -print STDERR "T_HDR_consist: $T_HDR_consist\n"; +print STDERR "T_HDR_rodata: $T_HDR_rodata\n"; print STDERR "T_HDR_closure: $T_HDR_closure\n"; print STDERR "T_HDR_info: $T_HDR_info\n"; print STDERR "T_HDR_entry: $T_HDR_entry\n"; print STDERR "T_HDR_vector: $T_HDR_vector\n"; -print STDERR "T_HDR_direct: $T_HDR_direct\n"; } } @@ -457,16 +551,55 @@ sub mangle_asm { if ( $TargetPlatform =~ /^mips-/ && /^\t\.(globl\S+\.text|comm\t)/ ) { $EXTERN_DECLS .= $_ unless /(__DISCARD__|\b(PK_|ASSIGN_)(FLT|DBL)\b)/; - - # As a temporary solution for compiling "foreign export" declarations, - # we use global variables to pass arguments from C to STG land. - # These declarations live in the .hc file and not in the generated C - # stub file, so we let them pass through here. - } elsif ( /^\t\.comm\t__fexp_.*$/ ) { + # Treat .comm variables as data. These show up in two (known) places: + # + # - the module_registered variable used in the __stginit fragment. + # even though these are declared static and initialised, gcc 3.3 + # likes to make them .comm, presumably to save space in the + # object file. + # + # - global variables used to pass arguments from C to STG in + # a foreign export. (is this still true? --SDM) + # + } elsif ( /^\t\.comm.*$/ ) { $chk[++$i] = $_; $chkcat[$i] = 'data'; $chksymb[$i] = ''; + # Labels ending "_str": these are literal strings. + } elsif ( /^${T_US}([A-Za-z0-9_]+)_str${T_POST_LBL}$/ ) { + $chk[++$i] = $_; + $chkcat[$i] = 'rodata'; + $chksymb[$i] = ''; + } elsif ( $TargetPlatform =~ /-darwin/ + && (/^\s*\.subsections_via_symbols/ + ||/^\s*\.no_dead_strip.*/)) { + # Don't allow Apple's linker to do any dead-stripping of symbols + # in this file, because it will mess up info-tables in mangled + # code. + # The .no_dead_strip directives are actually put there by + # the gcc3 "used" attribute on entry points. + + } elsif ( $TargetPlatform =~ /^.*-apple-darwin.*/ && ( + /^\s*\.picsymbol_stub/ + || /^\s*\.section __TEXT,__picsymbol_stub\d,.*/ + || /^\s*\.section __TEXT,__picsymbolstub\d,.*/ + || /^\s*\.symbol_stub/ + || /^\s*\.section __TEXT,__symbol_stub\d,.*/ + || /^\s*\.section __TEXT,__symbolstub\d,.*/ + || /^\s*\.lazy_symbol_pointer/ + || /^\s*\.non_lazy_symbol_pointer/ )) + { + $chk[++$i] = $_; + $chkcat[$i] = 'dyld'; + $chksymb[$i] = ''; + + } elsif ( $TargetPlatform =~ /^.*-apple-darwin.*/ && $chkcat[$i] eq 'dyld' && /^\s*\.data/) + { # non_lazy_symbol_ptrs that point to local symbols + $chk[++$i] = $_; + $chkcat[$i] = 'dyld'; + $chksymb[$i] = ''; + } elsif ( /^\s+/ ) { # most common case first -- a simple line! # duplicated from the bottom @@ -475,7 +608,6 @@ sub mangle_asm { } elsif ( /\.\.ng:$/ && $TargetPlatform =~ /^alpha-/ ) { # Alphas: Local labels not to be confused with new chunks $chk[$i] .= $_; - # NB: all the rest start with a non-space } elsif ( $TargetPlatform =~ /^mips-/ @@ -528,6 +660,11 @@ sub mangle_asm { $chkcat[$i] = 'data'; $chksymb[$i] = ''; + } elsif ( /^${T_US}(stg_ap_stack_entries|stg_stack_save_entries|stg_arg_bitmaps)${T_POST_LBL}$/o ) { + $chk[++$i] = $_; + $chkcat[$i] = 'data'; + $chksymb[$i] = ''; + } elsif ( /^(${T_US}__gnu_compiled_c|gcc2_compiled\.)${T_POST_LBL}/o ) { ; # toss it @@ -562,15 +699,6 @@ sub mangle_asm { $vectorchk{$1} = $i; - # As a temporary solution for compiling "foreign export" declarations, - # we use global variables to pass arguments from C to STG land. - # These declarations live in the .hc file and not in the generated C - # stub file, so we let them pass through here. - } elsif ( /^[\t ]+\.comm[\t ]+__fexp_.*$/ ) { - $chk[++$i] = $_; - $chkcat[$i] = 'data'; - $chksymb[$i] = ''; - } elsif ( $TargetPlatform =~ /^i386-.*-solaris2/ && /^[A-Za-z0-9][A-Za-z0-9_]*:/ ) { # Some Solaris system headers contain function definitions (as @@ -584,53 +712,59 @@ sub mangle_asm { $chkcat[$i] = 'misc'; $chksymb[$i] = $1; + } elsif ( $TargetPlatform =~ /^i386-apple-darwin/ && /^(___i686\.get_pc_thunk\.[abcd]x):/o) { + # To handle PIC on Darwin/x86, we need to appropriately pass through + # the get_pc_thunk functions. The need to be put into a special section + # marked as coalesced (otherwise the .weak_definition doesn't work + # on Darwin). + $chk[++$i] = $_; + $chkcat[$i] = 'get_pc_thunk'; + $chksymb[$i] = $1; + } elsif ( /^${T_US}[A-Za-z0-9_]/o && ( $TargetPlatform !~ /^hppa/ # need to avoid local labels in this case - || ! /^L\$\d+$/ ) ) { + || ! /^L\$\d+$/ ) + && ( $TargetPlatform !~ /^powerpc64/ # we need to avoid local labels in this case + || ! /^\.L\d+:$/ ) ) { local($thing); chop($thing = $_); $thing =~ s/:$//; - print STDERR "Warning: retaining unknown function \`$thing' in output from C compiler\n" - unless # $KNOWN_FUNNY_THING{$thing} + $chk[++$i] = $_; + $chksymb[$i] = ''; + if ( /^${T_US}stg_.*${T_POST_LBL}$/o # RTS internals || /^${T_US}__stg_.*${T_POST_LBL}$/o # more RTS internals || /^${T_US}__fexp_.*${T_POST_LBL}$/o # foreign export || /^${T_US}.*_slow${T_POST_LBL}$/o # slow entry || /^${T_US}__stginit.*${T_POST_LBL}$/o # __stginit || /^${T_US}.*_btm${T_POST_LBL}$/o # large bitmaps - || /^${T_US}.*_srtd${T_POST_LBL}$/o # large bitmaps || /^${T_US}.*_fast${T_POST_LBL}$/o # primops + || /^_uname:/o # x86/Solaris2 + ) + { + $chkcat[$i] = 'misc'; + } elsif ( + /^${T_US}.*_srtd${T_POST_LBL}$/o # large bitmaps || /^${T_US}.*_closure_tbl${T_POST_LBL}$/o # closure tables - || /^_uname:/o; # x86/Solaris2 - $chk[++$i] = $_; - $chkcat[$i] = 'misc'; - $chksymb[$i] = ''; + ) + { + $chkcat[$i] = 'relrodata'; + } else + { + print STDERR "Warning: retaining unknown function \`$thing' in output from C compiler\n"; + $chkcat[$i] = 'unknown'; + } - } elsif ( $TargetPlatform =~ /^powerpc-apple-.*/ && /^\.picsymbol_stub/ ) - { - $chk[++$i] = $_; - $chkcat[$i] = 'dyld'; - $chksymb[$i] = ''; - } elsif ( $TargetPlatform =~ /^powerpc-apple-.*/ && /^\.symbol_stub/ ) - { - $chk[++$i] = $_; - $chkcat[$i] = 'dyld'; - $chksymb[$i] = ''; - } elsif ( $TargetPlatform =~ /^powerpc-apple-.*/ && /^\.lazy_symbol_pointer/ ) - { - $chk[++$i] = $_; - $chkcat[$i] = 'dyld'; - $chksymb[$i] = ''; - } elsif ( $TargetPlatform =~ /^powerpc-apple-.*/ && /^\.non_lazy_symbol_pointer/ ) - { - $chk[++$i] = $_; - $chkcat[$i] = 'dyld'; - $chksymb[$i] = ''; - } elsif ( $TargetPlatform =~ /^powerpc-apple-.*/ && /^\.data/ && $chkcat[$i] eq 'dyld') - { # non_lazy_symbol_ptrs that point to local symbols + } elsif ( $TargetPlatform =~ /^powerpc-.*-linux/ && /^\.LCTOC1 = /o ) { + # PowerPC Linux's large-model PIC (-fPIC) generates a gobal offset + # table "by hand". Be sure to copy it over. + # Note that this label and all entries in the table should actually + # go into the .got2 section, but it isn't easy to distinguish them + # from other constant literals (.LC\d+), so we just put everything + # in .rodata. $chk[++$i] = $_; - $chkcat[$i] = 'dyld'; - $chksymb[$i] = ''; + $chkcat[$i] = 'literal'; + $chksymb[$i] = 'LCTOC1'; } else { # simple line (duplicated at the top) $chk[$i] .= $_; @@ -682,23 +816,52 @@ sub mangle_asm { $r =~ s/^\s*\n//; if ($TargetPlatform =~ /^i386-/) { - $p =~ s/^\tpushl\s+\%edi\n//; - $p =~ s/^\tpushl\s+\%esi\n//; - $p =~ s/^\tpushl\s+\%ebx\n//; - $p =~ s/^\tmovl\s+\%esi,\s*\d*\(\%esp\)\n//; - $p =~ s/^\tmovl\s+\%edi,\s*\d*\(\%esp\)\n//; - $p =~ s/^\tsubl\s+\$\d+,\s*\%esp\n//; - $p =~ s/^\tmovl\s+\$\d+,\s*\%eax\n\tcall\s+__alloca\n// if ($TargetPlatform =~ /^.*-cygwin32/); - - # GCC 3.1 is in the habit of adding spurious writes to the - # stack in the prologue. Just to be on the safe side, - # chuck these over the fence into the main code. - while ($p =~ /^\tmovl\s+\$\d+,\s*\d*\(\%esp\)\n/) { - # print "Spurious instruction: $&"; - $p = $` . $'; - $r = $& . $r; + if ($p =~ /^\tsubl\s+\$(\d+),\s*\%esp\n/) { + if ($1 >= 8192) { + die "Error: reserved stack space exceeded!\n Possible workarounds: compile with -fasm, or try another version of gcc.\n" + } } + # gcc 3.4.3 puts this kind of stuff in the prologue, eg. + # when compiling PrimOps.cmm with -optc-O2: + # xorl %ecx, %ecx + # xorl %edx, %edx + # movl %ecx, 16(%esp) + # movl %edx, 20(%esp) + # but then the code of the function doesn't assume + # anything about the contnets of these stack locations. + # I think it's to do with the use of inline functions for + # PK_Word64() and friends, where gcc is initialising the + # contents of the struct to zero, and failing to optimise + # away the initialisation. Let's live dangerously and + # discard these initalisations. + + $p =~ s/^\tpushl\s+\%e(di|si|bx)\n//g; + $p =~ s/^\txorl\s+\%e(ax|cx|dx),\s*\%e(ax|cx|dx)\n//g; + $p =~ s/^\tmovl\s+\%e(ax|cx|dx|si|di),\s*\d*\(\%esp\)\n//g; + $p =~ s/^\tmovl\s+\$\d+,\s*\d*\(\%esp\)\n//g; + $p =~ s/^\tsubl\s+\$\d+,\s*\%esp\n//; + $p =~ s/^\tmovl\s+\$\d+,\s*\%eax\n\tcall\s+__alloca\n// if ($TargetPlatform =~ /^.*-(cygwin32|mingw32)/); + + if ($TargetPlatform =~ /^i386-apple-darwin/) { + $pcrel_label = $p; + $pcrel_label =~ s/(.|\n)*^(\"?L\d+\$pb\"?):\n(.|\n)*/$2/ or $pcrel_label = ""; + $pcrel_reg = $p; + $pcrel_reg =~ s/(.|\n)*.*___i686\.get_pc_thunk\.([abcd]x)\n(.|\n)*/$2/ or $pcrel_reg = ""; + $p =~ s/^\s+call\s+___i686\.get_pc_thunk\..x//; + $p =~ s/^\"?L\d+\$pb\"?:\n//; + + if ($pcrel_reg eq "bx") { + # Bad gcc. Goes and uses %ebx, our BaseReg, for PIC. Bad gcc. + die "Darwin/x86: -fPIC -via-C doesn't work yet, use -fasm. Aborting." + } + } + + } elsif ($TargetPlatform =~ /^x86_64-/) { + $p =~ s/^\tpushq\s+\%r(bx|bp|12|13|14)\n//g; + $p =~ s/^\tmovq\s+\%r(bx|bp|12|13|14),\s*\d*\(\%rsp\)\n//g; + $p =~ s/^\tsubq\s+\$\d+,\s*\%rsp\n//; + } elsif ($TargetPlatform =~ /^ia64-/) { $p =~ s/^\t\.prologue .*\n//; $p =~ s/^\t\.save ar\.pfs, r\d+\n\talloc r\d+ = ar\.pfs, 0, 3[12], \d+, 0\n//; @@ -736,14 +899,14 @@ sub mangle_asm { $p =~ s/^\tsw\t\$fp,\d+\(\$sp\)\n//; $p =~ s/^\tsw\t\$28,\d+\(\$sp\)\n//; $p =~ s/__FRAME__/$FRAME/; - } elsif ($TargetPlatform =~ /^powerpc-apple-.*/) { + } elsif ($TargetPlatform =~ /^powerpc-apple-darwin.*/) { $pcrel_label = $p; - $pcrel_label =~ s/(.|\n)*^(L\d+\$pb):\n(.|\n)*/$2/ or $pcrel_label = ""; + $pcrel_label =~ s/(.|\n)*^(\"?L\d+\$pb\"?):\n(.|\n)*/$2/ or $pcrel_label = ""; $p =~ s/^\tmflr r0\n//; $p =~ s/^\tbl saveFP # f\d+\n//; $p =~ s/^\tbl saveFP ; save f\d+-f\d+\n//; - $p =~ s/^L\d+\$pb:\n//; + $p =~ s/^\"?L\d+\$pb\"?:\n//; $p =~ s/^\tstmw r\d+,-\d+\(r1\)\n//; $p =~ s/^\tstfd f\d+,-\d+\(r1\)\n//g; $p =~ s/^\tstw r0,\d+\(r1\)\n//g; @@ -758,6 +921,53 @@ sub mangle_asm { # I have no idea why, and I don't think it is necessary, so let's toss it. $p =~ s/^\tli r\d+,0\n//g; $p =~ s/^\tstw r\d+,\d+\(r1\)\n//g; + } elsif ($TargetPlatform =~ /^powerpc-.*-linux/) { + $p =~ s/^\tmflr 0\n//; + $p =~ s/^\tstmw \d+,\d+\(1\)\n//; + $p =~ s/^\tstfd \d+,\d+\(1\)\n//g; + $p =~ s/^\tstw r0,8\(1\)\n//; + $p =~ s/^\tstwu 1,-\d+\(1\)\n//; + $p =~ s/^\tstw \d+,\d+\(1\)\n//g; + + # GCC's "large-model" PIC (-fPIC) + $pcrel_label = $p; + $pcrel_label =~ s/(.|\n)*^.LCF(\d+):\n(.|\n)*/$2/ or $pcrel_label = ""; + + $p =~ s/^\tbcl 20,31,.LCF\d+\n//; + $p =~ s/^.LCF\d+:\n//; + $p =~ s/^\tmflr 30\n//; + $p =~ s/^\tlwz 0,\.LCL\d+-\.LCF\d+\(30\)\n//; + $p =~ s/^\tadd 30,0,30\n//; + + # This is bad: GCC 3 seems to zero-fill some local variables in the prologue + # under some circumstances, only when generating position dependent code. + # I have no idea why, and I don't think it is necessary, so let's toss it. + $p =~ s/^\tli \d+,0\n//g; + $p =~ s/^\tstw \d+,\d+\(1\)\n//g; + } elsif ($TargetPlatform =~ /^powerpc64-.*-linux/) { + $p =~ s/^\tmr 31,1\n//; + $p =~ s/^\tmflr 0\n//; + $p =~ s/^\tstmw \d+,\d+\(1\)\n//; + $p =~ s/^\tstfd \d+,-?\d+\(1\)\n//g; + $p =~ s/^\tstd r0,8\(1\)\n//; + $p =~ s/^\tstdu 1,-\d+\(1\)\n//; + $p =~ s/^\tstd \d+,-?\d+\(1\)\n//g; + + # GCC's "large-model" PIC (-fPIC) + $pcrel_label = $p; + $pcrel_label =~ s/(.|\n)*^.LCF(\d+):\n(.|\n)*/$2/ or $pcrel_label = ""; + + $p =~ s/^\tbcl 20,31,.LCF\d+\n//; + $p =~ s/^.LCF\d+:\n//; + $p =~ s/^\tmflr 30\n//; + $p =~ s/^\tlwz 0,\.LCL\d+-\.LCF\d+\(30\)\n//; + $p =~ s/^\tadd 30,0,30\n//; + + # This is bad: GCC 3 seems to zero-fill some local variables in the prologue + # under some circumstances, only when generating position dependent code. + # I have no idea why, and I don't think it is necessary, so let's toss it. + $p =~ s/^\tli \d+,0\n//g; + $p =~ s/^\tstd \d+,\d+\(1\)\n//g; } else { print STDERR "$Pgm: unknown prologue mangling? $TargetPlatform\n"; } @@ -766,13 +976,23 @@ sub mangle_asm { #print stderr "HWL: this should die! Prologue junk?: $p\n" if $p =~ /^\t[^\.]/; die "Prologue junk?: $p\n" if $p =~ /^\s+[^\s\.]/; - if ($TargetPlatform =~ /^powerpc-apple-.*/ && $pcrel_label ne "") { - # on PowerPC, we have to keep a part of the prologue - # (which loads the current instruction pointer into register r31) + # For PIC, we want to keep part of the prologue + if ($TargetPlatform =~ /^powerpc-apple-darwin.*/ && $pcrel_label ne "") { + # Darwin: load the current instruction pointer into register r31 $p .= "bcl 20,31,$pcrel_label\n"; $p .= "$pcrel_label:\n"; $p .= "\tmflr r31\n"; - } + } elsif ($TargetPlatform =~ /^powerpc-.*-linux/ && $pcrel_label ne "") { + # Linux: load the GOT pointer into register 30 + $p .= "\tbcl 20,31,.LCF$pcrel_label\n"; + $p .= ".LCF$pcrel_label:\n"; + $p .= "\tmflr 30\n"; + $p .= "\tlwz 0,.LCL$pcrel_label-.LCF$pcrel_label(30)\n"; + $p .= "\tadd 30,0,30\n"; + } elsif ($TargetPlatform =~ /^i386-apple-darwin.*/ && $pcrel_label ne "") { + $p .= "\tcall ___i686.get_pc_thunk.$pcrel_reg\n"; + $p .= "$pcrel_label:\n"; + } # glue together what's left $c = $p . $r; @@ -811,12 +1031,28 @@ sub mangle_asm { $e =~ s/^\tlw\t\$fp,\d+\(\$sp\)\n//; $e =~ s/^\taddu\t\$sp,\$sp,\d+\n//; $e =~ s/^\tj\t\$31\n//; - } elsif ($TargetPlatform =~ /^powerpc-apple-.*/) { + } elsif ($TargetPlatform =~ /^powerpc-apple-darwin.*/) { $e =~ s/^\taddi r1,r1,\d+\n//; - $e =~ s/^\tcal r1,\d+\(r1\)\n//; - $e =~ s/^\tlw?z? r\d+,\d+\(r1\)\n//; + $e =~ s/^\tlwz r\d+,\d+\(r1\)\n//; + $e =~ s/^\tlmw r\d+,-\d+\(r1\)\n//; $e =~ s/^\tmtlr r0\n//; $e =~ s/^\tblr\n//; + $e =~ s/^\tb restFP ;.*\n//; + } elsif ($TargetPlatform =~ /^powerpc64-.*-linux/) { + $e =~ s/^\tmr 3,0\n//; + $e =~ s/^\taddi 1,1,\d+\n//; + $e =~ s/^\tld 0,16\(1\)\n//; + $e =~ s/^\tmtlr 0\n//; + + # callee-save registers + $e =~ s/^\tld \d+,-?\d+\(1\)\n//g; + $e =~ s/^\tlfd \d+,-?\d+\(1\)\n//g; + + # get rid of the debug junk along with the blr + $e =~ s/^\tblr\n\t.long .*\n\t.byte .*\n//; + + # incase we missed it with the last one get the blr alone + $e =~ s/^\tblr\n//; } else { print STDERR "$Pgm: unknown epilogue mangling? $TargetPlatform\n"; } @@ -831,9 +1067,12 @@ sub mangle_asm { # On SPARCs, we don't do --- BEGIN/END ---, we just # toss the register-windowing save/restore/ret* instructions - # directly: + # directly unless they've been generated by function definitions in header + # files on Solaris: if ( $TargetPlatform =~ /^sparc-/ ) { - $c =~ s/^\t(save.*|restore.*|ret|retl)\n//g; + if ( ! ( $TargetPlatform =~ /solaris2$/ && $chkcat[$i] eq 'unknown' )) { + $c =~ s/^\t(save.*|restore.*|ret|retl)\n//g; + } # throw away PROLOGUE comments $c =~ s/^\t!#PROLOGUE# 0\n\t!#PROLOGUE# 1\n//; } @@ -843,7 +1082,10 @@ sub mangle_asm { # toss all calls to __DISCARD__ $c =~ s/^\t(call|jbsr|jal)\s+${T_US}__DISCARD__\n//go; $c =~ s/^\tjsr\s+\$26\s*,\s*${T_US}__DISCARD__\n//go if $TargetPlatform =~ /^alpha-/; - $c =~ s/^\tbl\s+L___DISCARD__\$stub\n//go if $TargetPlatform =~ /^powerpc-apple-.*/; + $c =~ s/^\tbl\s+L___DISCARD__\$stub\n//go if $TargetPlatform =~ /^powerpc-apple-darwin.*/; + $c =~ s/^\tbl\s+__DISCARD__(\@plt)?\n//go if $TargetPlatform =~ /^powerpc-.*-linux/; + $c =~ s/^\tbl\s+\.__DISCARD__\n\s+nop\n//go if $TargetPlatform =~ /^powerpc64-.*-linux/; + $c =~ s/^\tcall\s+L___DISCARD__\$stub\n//go if $TargetPlatform =~ /i386-apple-darwin.*/; # IA64: mangle tailcalls into jumps here if ($TargetPlatform =~ /^ia64-/) { @@ -971,7 +1213,23 @@ sub mangle_asm { # print out all the literal strings next for ($i = 0; $i < $numchks; $i++) { if ( $chkcat[$i] eq 'literal' ) { - print OUTASM $T_HDR_literal, $chk[$i]; + + # HACK: try to detect 16-byte constants and align them + # on a 16-byte boundary. x86_64 sometimes needs 128-bit + # aligned constants, and so does Darwin/x86. + if ( $TargetPlatform =~ /^x86_64/ + || $TargetPlatform =~ /^i386-apple-darwin/ ) { + $z = $chk[$i]; + if ($z =~ /(\.long.*\n.*\.long.*\n.*\.long.*\n.*\.long|\.quad.*\n.*\.quad)/) { + print OUTASM $T_HDR_literal16; + } else { + print OUTASM $T_HDR_literal; + } + } else { + print OUTASM $T_HDR_literal; + } + + print OUTASM $chk[$i]; print OUTASM "; end literal\n" if $TargetPlatform =~ /^hppa/; # for the splitter $chkcat[$i] = 'DONE ALREADY'; @@ -995,7 +1253,7 @@ sub mangle_asm { next if $chkcat[$i] eq 'DONE ALREADY'; - if ( $chkcat[$i] eq 'misc' ) { + if ( $chkcat[$i] eq 'misc' || $chkcat[$i] eq 'unknown' ) { if ($chk[$i] ne '') { print OUTASM $T_HDR_misc; &print_doctored($chk[$i], 0); @@ -1010,42 +1268,6 @@ sub mangle_asm { print OUTASM $chk[$i]; } - } elsif ( $chkcat[$i] eq 'consist' ) { - if ( $chk[$i] =~ /$T_hsc_cc_PAT/o ) { - local($consist) = "$1.$2.$3"; - $consist =~ s/,/./g; - $consist =~ s/\//./g; - $consist =~ s/-/_/g; - $consist =~ s/[^A-Za-z0-9_.]/ZZ/g; # ToDo: properly? - # - # Using a cygnus-2.7-96q4 gcc build on hppas, the - # consistency chunk for ghc_cc_ID often (but not always!) - # gets lumped with a bunch of .IMPORT directives containing info on - # the code or data space nature of external symbols. We can't - # toss these, so once the consistency ID has been turned into - # a representable symbol, we substitute it for the symbol - # that the string was attached to in the first place (ghc_cc_ID.) - # (The original string is also substituted away.) - # - # This change may affect the code output on other platforms in - # adverse ways, hence we restrict this hack hppa targets only. - # - # -- 2/98 SOF - if ( $TargetPlatform =~ /^hppa/ ) { - $chk[$i] =~ s/^${T_US}ghc.*c_ID$TPOSTLBL/$consist/o; - $chk[$i] =~ s/\t$T_hsc_cc_PAT/$T_HDR_misc/o; - $consist = $chk[$i]; #clumsily - } - print OUTASM $T_HDR_consist, "${consist}${T_POST_LBL}\n"; - - } elsif ( $TargetPlatform !~ /^(mips)-/ ) { # we just don't try in those case (ToDo) - # on mips: consistency string is just a v - # horrible bunch of .bytes, - # which I am too lazy to sort out (WDP 95/05) - - print STDERR "Couldn't grok consistency: ", $chk[$i]; - } - } elsif ( $chkcat[$i] eq 'splitmarker' ) { # we can just re-constitute this one... # NB: we emit _three_ underscores no matter what, @@ -1067,7 +1289,7 @@ sub mangle_asm { # SRT if ( defined($srtchk{$symb}) ) { - print OUTASM $T_HDR_srt; + print OUTASM $T_HDR_relrodata; print OUTASM $chk[$srtchk{$symb}]; $chkcat[$srtchk{$symb}] = 'DONE ALREADY'; } @@ -1083,11 +1305,33 @@ sub mangle_asm { $chkcat[$infochk{$symb}] = 'DONE ALREADY'; } - # STD ENTRY POINT + # ENTRY POINT if ( defined($entrychk{$symb}) ) { $c = $chk[$entrychk{$symb}]; + # If this is an entry point with an info table, + # eliminate the entry symbol and all directives involving it. + if (defined($infochk{$symb}) && $TargetPlatform !~ /^ia64-/) { + @o = (); + foreach $l (split(/\n/,$c)) { + next if $l =~ /^.*$symb_(entry|ret)${T_POST_LBL}/; + + # If we have .type/.size direrctives involving foo_entry, + # then make them refer to foo_info instead. The information + # in these directives is used by the cachegrind annotator, + # so it is worthwhile keeping. + if ($l =~ /^\s*\.(type|size).*$symb_(entry|ret)/) { + $l =~ s/$symb(_entry|_ret)/${symb}_info/g; + push(@o,$l); + next; + } + next if $l =~ /^\s*\..*$symb.*\n?/; + push(@o,$l); + } + $c = join("\n",@o) . "\n"; + } + print OUTASM $T_HDR_entry; &print_doctored($c, 1); # NB: the 1!!! @@ -1095,8 +1339,7 @@ sub mangle_asm { $chkcat[$entrychk{$symb}] = 'DONE ALREADY'; } - } elsif ( $chkcat[$i] eq 'vector' - || $chkcat[$i] eq 'direct' ) { # do them in that order + } elsif ( $chkcat[$i] eq 'vector' ) { $symb = $chksymb[$i]; # VECTOR TABLE @@ -1117,6 +1360,14 @@ sub mangle_asm { print OUTASM "\t# nop\n"; } + } elsif ( $chkcat[$i] eq 'rodata' ) { + print OUTASM $T_HDR_rodata; + print OUTASM $chk[$i]; + $chkcat[$i] = 'DONE ALREADY'; + } elsif ( $chkcat[$i] eq 'relrodata' ) { + print OUTASM $T_HDR_relrodata; + print OUTASM $chk[$i]; + $chkcat[$i] = 'DONE ALREADY'; } elsif ( $chkcat[$i] eq 'toc' ) { # silly optimisation to print tocs, since they come in groups... print OUTASM $T_HDR_toc; @@ -1131,12 +1382,16 @@ sub mangle_asm { $j++; } - } elsif ( $TargetPlatform =~ /^powerpc-apple-.*/ && $chkcat[$i] eq 'dyld' ) { - # powerpc-apple: dynamic linker stubs + } elsif ( $TargetPlatform =~ /^.*-apple-darwin.*/ && $chkcat[$i] eq 'dyld' ) { + # apple-darwin: dynamic linker stubs if($chk[$i] !~ /\.indirect_symbol ___DISCARD__/) { # print them out unchanged, but remove the stubs for __DISCARD__ print OUTASM $chk[$i]; } + } elsif ( $TargetPlatform =~ /^i386-apple-darwin.*/ && $chkcat[$i] eq 'get_pc_thunk' ) { + # i386-apple-darwin: __i686.get_pc_thunk.[abcd]x + print OUTASM ".section __TEXT,__textcoal_nt,coalesced,no_toc\n"; + print OUTASM $chk[$i]; } else { &tidy_up_and_die(1,"$Pgm: unknown chkcat (ghc-asm: $TargetPlatform)\n$chkcat[$i]\n$chk[$i]\n"); } @@ -1179,12 +1434,25 @@ sub hppa_mash_prologue { # OK, epilogue, too sub print_doctored { local($_, $need_fallthru_patch) = @_; + if ( $TargetPlatform =~ /^x86_64-/ ) { + # Catch things like + # + # movq -4(%ebp), %rax + # jmp *%rax + # + # and optimise: + # + s/^\tmovq\s+(-?\d*\(\%r(bx|bp|13)\)),\s*(\%r(ax|cx|dx|10|11))\n\tjmp\s+\*\3/\tjmp\t\*$1/g; + s/^\tmovl\s+\$${T_US}(.*),\s*(\%e(ax|cx|si|di))\n\tjmp\s+\*\%r\3/\tjmp\t$T_US$1/g; + } + if ( $TargetPlatform !~ /^i386-/ || ! /^\t[a-z]/ # no instructions in here, apparently || /^${T_US}__stginit_[A-Za-z0-9_]+${T_POST_LBL}/) { print OUTASM $_; return; } + # OK, must do some x86 **HACKING** local($entry_patch) = ''; @@ -1223,8 +1491,16 @@ sub print_doctored { # movl $_blah, # jmp * # + s/^\tmovl\s+\$${T_US}(.*),\s*(\%e[acd]x)\n\tjmp\s+\*\2/\tjmp $T_US$1/g; - s/^\tmovl\s+\$${T_US}(.*),\s*(\%e[abcd]x)\n\tjmp\s+\*\2/\tjmp $T_US$1/g; + # Catch things like + # + # movl -4(%ebx), %eax + # jmp *%eax + # + # and optimise: + # + s/^\tmovl\s+(-?\d*\(\%e(bx|si)\)),\s*(\%e[acd]x)\n\tjmp\s+\*\3/\tjmp\t\*$1/g; if ($StolenX86Regs <= 2 ) { # YURGH! spurious uses of esi? s/^\tmovl\s+(.*),\s*\%esi\n\tjmp\s+\*%esi\n/\tmovl $1,\%eax\n\tjmp \*\%eax\n/g; @@ -1275,6 +1551,16 @@ sub print_doctored { # fix _all_ non-local jumps: + if ( $TargetPlatform =~ /^.*-apple-darwin.*/ ) { + # On Darwin, we've got local-looking jumps that are + # actually global (i.e. jumps to Lfoo$stub or via + # Lfoo$non_lazy_ptr), so we fix those first. + # In fact, we just fix everything that contains a dollar + # because false positives don't hurt here. + + s/^(\tjmp\s+\*?L.*\$.*\n)/$exit_patch$1/g; + } + s/^\tjmp\s+\*${T_X86_PRE_LLBL_PAT}/\tJMP___SL/go; s/^\tjmp\s+${T_X86_PRE_LLBL_PAT}/\tJMP___L/go; @@ -1344,13 +1630,30 @@ sub rev_tbl { $before .= $lines[$i] . "\n"; # otherwise... } + $infoname = $label; + $infoname =~ s/(.|\n)*^([A-Za-z0-9_]+_info)${T_POST_LBL}$(.|\n)*/\2/; + # Grab the table data... if ( $TargetPlatform !~ /^hppa/ ) { for ( ; $i <= $#lines && $lines[$i] =~ /^\t?${T_DOT_WORD}\s+/o; $i++) { - push(@words, $lines[$i]); + $line = $lines[$i]; + # Convert addresses of SRTs, slow entrypoints and large bitmaps + # to offsets (relative to the info label), + # in order to support position independent code. + $line =~ s/$infoname/0/ + || $line =~ s/([A-Za-z0-9_]+_srtd)$/\1 - $infoname/ + || $line =~ s/([A-Za-z0-9_]+_srt(\+\d+)?)$/\1 - $infoname/ + || $line =~ s/([A-Za-z0-9_]+_slow)$/\1 - $infoname/ + || $line =~ s/([A-Za-z0-9_]+_btm)$/\1 - $infoname/ + || $line =~ s/([A-Za-z0-9_]+_alt)$/\1 - $infoname/ + || $line =~ s/([A-Za-z0-9_]+_dflt)$/\1 - $infoname/ + || $line =~ s/([A-Za-z0-9_]+_ret)$/\1 - $infoname/; + push(@words, $line); } } else { # hppa weirdness for ( ; $i <= $#lines && $lines[$i] =~ /^\s+(${T_DOT_WORD}|\.IMPORT)/; $i++) { + # FIXME: the RTS now expects offsets instead of addresses + # for all labels in info tables. if ($lines[$i] =~ /^\s+\.IMPORT/) { push(@imports, $lines[$i]); } else {