maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/test/CodeGen/X86/global-sections.ll

302 lines
7.6 KiB
LLVM
Raw Normal View History

Eliminate more uses of llvm-as and llvm-dis. llvm-svn: 81290
2009-09-09 07:54:48 +08:00
; RUN: llc < %s -mtriple=i386-unknown-linux-gnu | FileCheck %s -check-prefix=LINUX
; RUN: llc < %s -mtriple=i386-apple-darwin9.7 | FileCheck %s -check-prefix=DARWIN
Use __literal16. It has been supported by the linker since 2005. llvm-svn: 201365
2014-02-14 07:16:11 +08:00
; RUN: llc < %s -mtriple=i386-apple-darwin10 -relocation-model=static | FileCheck %s -check-prefix=DARWIN-STATIC
; RUN: llc < %s -mtriple=x86_64-apple-darwin10 | FileCheck %s -check-prefix=DARWIN64
Put each jump table in an independent section if the function is too. This allows the linker to GC both, fixing pr22557. llvm-svn: 228937
2015-02-13 01:16:46 +08:00
; RUN: llc < %s -mtriple=i386-unknown-linux-gnu -data-sections -function-sections | FileCheck %s -check-prefix=LINUX-SECTIONS
Put jump tables in distinct sections if -ffunction-sections is used. A small regression in r230411 was that we were basing the decision on -fdata-sections. llvm-svn: 230707
2015-02-27 07:55:11 +08:00
; RUN: llc < %s -mtriple=i386-unknown-linux-gnu -function-sections | FileCheck %s -check-prefix=LINUX-FUNC-SECTIONS
Add r228939 back with a fix. The problem in the original patch was not switching back to .text after printing an eh table. Original message: On ELF, put PIC jump tables in a non executable section. Fixes PR22558. llvm-svn: 229586
2015-02-18 07:34:51 +08:00
; RUN: llc < %s -mtriple=x86_64-pc-linux -data-sections -function-sections -relocation-model=pic | FileCheck %s -check-prefix=LINUX-SECTIONS-PIC
Move the function and data section flags into the options struct and make the functions to set them non-static. Move and rename the llvm specific backend options to avoid conflicting with the clang option. Paired with a backend commit to update. llvm-svn: 209238
2014-05-21 05:25:34 +08:00
; RUN: llc < %s -mtriple=i686-pc-win32 -data-sections -function-sections | FileCheck %s -check-prefix=WIN32-SECTIONS
Put jump tables in unique sections on COFF. If a function is going in an unique section (because of -ffunction-sections for example), putting a jump table in .rodata will keep .rodata alive and that will keep alive any other function that also has a jump table. Instead, put the jump table in a unique section that is associated with the function. llvm-svn: 231961
2015-03-12 03:58:37 +08:00
; RUN: llc < %s -mtriple=i686-pc-win32 -function-sections | FileCheck %s -check-prefix=WIN32-FUNC-SECTIONS
add the most expedient hack to fix PR4619, along with a testcase. Thanks to Rafael for the great example. llvm-svn: 77083
2009-07-26 01:57:37 +08:00
WinCOFF: Add support for -fdata-sections This is a pretty straight forward translation for COFF, we just need to stick the data in a COMDAT section marked as IMAGE_COMDAT_SELECT_NODUPLICATES. N.B. We must be careful to avoid sticking entities with private linkage in COMDAT groups. COFF is pretty hostile to the renaming of entities so we must be careful to disallow GlobalVariables with unstable names. llvm-svn: 204703
2014-03-25 14:14:26 +08:00
define void @F1() {
ret void
}
; WIN32-SECTIONS: .section .text,"xr",one_only,_F1
; WIN32-SECTIONS: .globl _F1
simplify DarwinTargetAsmInfo::SelectSectionForGlobal a bit and make it more aggressive, we now put: const int G2 __attribute__((weak)) = 42; into the text (readonly) segment like gcc, previously we put it into the data (readwrite) segment. llvm-svn: 77104
2009-07-26 08:51:36 +08:00
Put each jump table in an independent section if the function is too. This allows the linker to GC both, fixing pr22557. llvm-svn: 228937
2015-02-13 01:16:46 +08:00
define void @F2(i32 %y) {
bb0:
switch i32 %y, label %bb5 [
i32 1, label %bb1
i32 2, label %bb2
i32 3, label %bb3
i32 4, label %bb4
]
bb1:
ret void
bb2:
ret void
bb3:
ret void
bb4:
ret void
bb5:
ret void
}
; LINUX: .size F2,
; LINUX-NEX: .cfi_endproc
; LINUX-NEX: .section .rodata,"a",@progbits
; LINUX-SECTIONS: .section .text.F2,"ax",@progbits
; LINUX-SECTIONS: .size F2,
; LINUX-SECTIONS-NEXT: .cfi_endproc
; LINUX-SECTIONS-NEXT: .section .rodata.F2,"a",@progbits
Put jump tables in distinct sections if -ffunction-sections is used. A small regression in r230411 was that we were basing the decision on -fdata-sections. llvm-svn: 230707
2015-02-27 07:55:11 +08:00
; LINUX-FUNC-SECTIONS: .section .text.F2,"ax",@progbits
; LINUX-FUNC-SECTIONS: .size F2,
; LINUX-FUNC-SECTIONS-NEXT: .cfi_endproc
; LINUX-FUNC-SECTIONS-NEXT: .section .rodata.F2,"a",@progbits
Put jump tables in unique sections on COFF. If a function is going in an unique section (because of -ffunction-sections for example), putting a jump table in .rodata will keep .rodata alive and that will keep alive any other function that also has a jump table. Instead, put the jump table in a unique section that is associated with the function. llvm-svn: 231961
2015-03-12 03:58:37 +08:00
; WIN32-FUNC-SECTIONS: .section .text,"xr",one_only,_F2
; WIN32-FUNC-SECTIONS-NOT: .section
; WIN32-FUNC-SECTIONS: .section .rdata,"dr",associative,_F2
Add r228939 back with a fix. The problem in the original patch was not switching back to .text after printing an eh table. Original message: On ELF, put PIC jump tables in a non executable section. Fixes PR22558. llvm-svn: 229586
2015-02-18 07:34:51 +08:00
; LINUX-SECTIONS-PIC: .section .text.F2,"ax",@progbits
; LINUX-SECTIONS-PIC: .size F2,
; LINUX-SECTIONS-PIC-NEXT: .cfi_endproc
; LINUX-SECTIONS-PIC-NEXT: .section .rodata.F2,"a",@progbits
Add a test showing the problem in r228939. If an EH table is printed in between the function and the jump table we would fail to switch back to the text section to print the jump table. llvm-svn: 229580
2015-02-18 07:21:46 +08:00
declare void @G()
define void @F3(i32 %y) {
bb0:
invoke void @G()
to label %bb2 unwind label %bb1
bb1:
landingpad { i8*, i32 } personality i8* bitcast (void ()* @G to i8*)
catch i8* null
br label %bb2
bb2:
switch i32 %y, label %bb7 [
i32 1, label %bb3
i32 2, label %bb4
i32 3, label %bb5
i32 4, label %bb6
]
bb3:
ret void
bb4:
ret void
bb5:
ret void
bb6:
ret void
bb7:
ret void
}
; DARWIN64: _F3:
Drop the "eh_" from eh_func_begin and eh_func_end. They will be used for more than eh tables. llvm-svn: 231185
2015-03-04 08:27:43 +08:00
; DARWIN64: Lfunc_end
Centralize handling of the eh_begin and eh_end labels. This removes a bit of duplicated code and more importantly, remembers the labels so that they don't need to be looked up by name. This in turn allows for any name to be used and avoids a crash if the name we wanted was already taken. llvm-svn: 230772
2015-02-28 02:18:39 +08:00
; DARWIN64-NEXT: .cfi_endproc
Add a test showing the problem in r228939. If an EH table is printed in between the function and the jump table we would fail to switch back to the text section to print the jump table. llvm-svn: 229580
2015-02-18 07:21:46 +08:00
; DARWIN64-NOT: .section
; DARWIN64: LJTI{{.*}}:
Print jump tables before exception tables. In the case where just tables are part of the function section, this produces more readable assembly by avoiding switching to the eh section and back to .text. This would also break with non unique section names, as trying to switch to a unique section actually creates a new one. llvm-svn: 231677
2015-03-10 02:29:12 +08:00
; DARWIN64-NEXT: .long
; DARWIN64-NEXT: .long
; DARWIN64-NEXT: .long
; DARWIN64-NEXT: .long
; DARWIN64-NEXT: .section __TEXT,__gcc_except_tab
Add a test showing the problem in r228939. If an EH table is printed in between the function and the jump table we would fail to switch back to the text section to print the jump table. llvm-svn: 229580
2015-02-18 07:21:46 +08:00
simplify DarwinTargetAsmInfo::SelectSectionForGlobal a bit and make it more aggressive, we now put: const int G2 __attribute__((weak)) = 42; into the text (readonly) segment like gcc, previously we put it into the data (readwrite) segment. llvm-svn: 77104
2009-07-26 08:51:36 +08:00
; int G1;
add the most expedient hack to fix PR4619, along with a testcase. Thanks to Rafael for the great example. llvm-svn: 77083
2009-07-26 01:57:37 +08:00
@G1 = common global i32 0
; LINUX: .type G1,@object
; LINUX: .comm G1,4,4
Move the function and data section flags into the options struct and make the functions to set them non-static. Move and rename the llvm specific backend options to avoid conflicting with the clang option. Paired with a backend commit to update. llvm-svn: 209238
2014-05-21 05:25:34 +08:00
; DARWIN: .comm _G1,4,2
simplify DarwinTargetAsmInfo::SelectSectionForGlobal a bit and make it more aggressive, we now put: const int G2 __attribute__((weak)) = 42; into the text (readonly) segment like gcc, previously we put it into the data (readwrite) segment. llvm-svn: 77104
2009-07-26 08:51:36 +08:00
; const int G2 __attribute__((weak)) = 42;
Move the function and data section flags into the options struct and make the functions to set them non-static. Move and rename the llvm specific backend options to avoid conflicting with the clang option. Paired with a backend commit to update. llvm-svn: 209238
2014-05-21 05:25:34 +08:00
@G2 = weak_odr unnamed_addr constant i32 42
simplify DarwinTargetAsmInfo::SelectSectionForGlobal a bit and make it more aggressive, we now put: const int G2 __attribute__((weak)) = 42; into the text (readonly) segment like gcc, previously we put it into the data (readwrite) segment. llvm-svn: 77104
2009-07-26 08:51:36 +08:00
; TODO: linux drops this into .rodata, we drop it into ".gnu.linkonce.r.G2"
Revert "Reapply r185872 now that the address sanitizer has been changed to support this." This reverts commit r187939. It broke an O0 build of a spec benchmark. llvm-svn: 188012
2013-08-09 05:04:16 +08:00
; DARWIN: .section __TEXT,__const_coal,coalesced
simplify DarwinTargetAsmInfo::SelectSectionForGlobal a bit and make it more aggressive, we now put: const int G2 __attribute__((weak)) = 42; into the text (readonly) segment like gcc, previously we put it into the data (readwrite) segment. llvm-svn: 77104
2009-07-26 08:51:36 +08:00
; DARWIN: _G2:
; DARWIN: .long 42
finish simplifying DarwinTargetAsmInfo::SelectSectionForGlobal for now. Make the section switching directives more consistent by not including \n and including \t for them all. llvm-svn: 77107
2009-07-26 09:24:18 +08:00
; int * const G3 = &G1;
Update tests. llvm-svn: 123591
2011-01-17 02:02:57 +08:00
@G3 = unnamed_addr constant i32* @G1
finish simplifying DarwinTargetAsmInfo::SelectSectionForGlobal for now. Make the section switching directives more consistent by not including \n and including \t for them all. llvm-svn: 77107
2009-07-26 09:24:18 +08:00
Make the big switch: Change MCSectionMachO to represent a section *semantically* instead of syntactically as a string. This means that it keeps track of the segment, section, flags, etc directly and asmprints them in the right format. This also includes parsing and validation support for llvm-mc and "attribute(section)", so we should now start getting errors about invalid section attributes from the compiler instead of the assembler on darwin. Still todo: 1) Uniquing of darwin mcsections 2) Move all the Darwin stuff out to MCSectionMachO.[cpp|h] 3) there are a few FIXMEs, for example what is the syntax to get the S_GB_ZEROFILL segment type? llvm-svn: 78547
2009-08-10 09:39:42 +08:00
; DARWIN: .section __DATA,__const
finish simplifying DarwinTargetAsmInfo::SelectSectionForGlobal for now. Make the section switching directives more consistent by not including \n and including \t for them all. llvm-svn: 77107
2009-07-26 09:24:18 +08:00
; DARWIN: .globl _G3
; DARWIN: _G3:
; DARWIN: .long _G1
Be nice to Xcore and the XMOS assembler and avoid quoting section names that contain only letters, digits and the characters "_" and ".". llvm-svn: 127028
2011-03-05 04:03:14 +08:00
; LINUX: .section .rodata,"a",@progbits
add llvm codegen support for -ffunction-sections and -fdata-sections, patch by Sylvere Teissier! llvm-svn: 101106
2010-04-13 08:36:43 +08:00
; LINUX: .globl G3
Be nice to Xcore and the XMOS assembler and avoid quoting section names that contain only letters, digits and the characters "_" and ".". llvm-svn: 127028
2011-03-05 04:03:14 +08:00
; LINUX-SECTIONS: .section .rodata.G3,"a",@progbits
add llvm codegen support for -ffunction-sections and -fdata-sections, patch by Sylvere Teissier! llvm-svn: 101106
2010-04-13 08:36:43 +08:00
; LINUX-SECTIONS: .globl G3
MC: Emit COFF section flags in the "proper" order COFF section flags are not idempotent: 'rd' will make a read-write section because 'd' implies write 'dr' will make a read-only section because 'r' disables write llvm-svn: 228490
2015-02-07 16:26:40 +08:00
; WIN32-SECTIONS: .section .rdata,"dr",one_only,_G3
WinCOFF: Add support for -fdata-sections This is a pretty straight forward translation for COFF, we just need to stick the data in a COMDAT section marked as IMAGE_COMDAT_SELECT_NODUPLICATES. N.B. We must be careful to avoid sticking entities with private linkage in COMDAT groups. COFF is pretty hostile to the renaming of entities so we must be careful to disallow GlobalVariables with unstable names. llvm-svn: 204703
2014-03-25 14:14:26 +08:00
; WIN32-SECTIONS: .globl _G3
finish simplifying DarwinTargetAsmInfo::SelectSectionForGlobal for now. Make the section switching directives more consistent by not including \n and including \t for them all. llvm-svn: 77107
2009-07-26 09:24:18 +08:00
; _Complex long long const G4 = 34;
On MachO, don't put non-private constants in mergeable sections. On MachO, putting a symbol that doesn't start with a 'L' or 'l' in one of the __TEXT,__literal* sections prevents the linker from merging the context of the section. Since private GVs are the ones the get mangled to start with 'L' or 'l', we now only put those on the __TEXT,__literal* sections. llvm-svn: 216682
2014-08-29 04:13:31 +08:00
@G4 = private unnamed_addr constant {i64,i64} { i64 34, i64 0 }
finish simplifying DarwinTargetAsmInfo::SelectSectionForGlobal for now. Make the section switching directives more consistent by not including \n and including \t for them all. llvm-svn: 77107
2009-07-26 09:24:18 +08:00
Fix an obvious typo that's preventing x86 (32-bit) from using .literal16. llvm-svn: 135535
2011-07-20 07:14:32 +08:00
; DARWIN: .section __TEXT,__literal16,16byte_literals
On MachO, don't put non-private constants in mergeable sections. On MachO, putting a symbol that doesn't start with a 'L' or 'l' in one of the __TEXT,__literal* sections prevents the linker from merging the context of the section. Since private GVs are the ones the get mangled to start with 'L' or 'l', we now only put those on the __TEXT,__literal* sections. llvm-svn: 216682
2014-08-29 04:13:31 +08:00
; DARWIN: L_G4:
put normal data into .data instead of .data.rel on elf systems. llvm-svn: 77116
2009-07-26 11:06:11 +08:00
; DARWIN: .long 34
Use __literal16. It has been supported by the linker since 2005. llvm-svn: 201365
2014-02-14 07:16:11 +08:00
; DARWIN-STATIC: .section __TEXT,__literal16,16byte_literals
On MachO, don't put non-private constants in mergeable sections. On MachO, putting a symbol that doesn't start with a 'L' or 'l' in one of the __TEXT,__literal* sections prevents the linker from merging the context of the section. Since private GVs are the ones the get mangled to start with 'L' or 'l', we now only put those on the __TEXT,__literal* sections. llvm-svn: 216682
2014-08-29 04:13:31 +08:00
; DARWIN-STATIC: L_G4:
Use __literal16. It has been supported by the linker since 2005. llvm-svn: 201365
2014-02-14 07:16:11 +08:00
; DARWIN-STATIC: .long 34
; DARWIN64: .section __TEXT,__literal16,16byte_literals
On MachO, don't put non-private constants in mergeable sections. On MachO, putting a symbol that doesn't start with a 'L' or 'l' in one of the __TEXT,__literal* sections prevents the linker from merging the context of the section. Since private GVs are the ones the get mangled to start with 'L' or 'l', we now only put those on the __TEXT,__literal* sections. llvm-svn: 216682
2014-08-29 04:13:31 +08:00
; DARWIN64: L_G4:
Use __literal16. It has been supported by the linker since 2005. llvm-svn: 201365
2014-02-14 07:16:11 +08:00
; DARWIN64: .quad 34
put normal data into .data instead of .data.rel on elf systems. llvm-svn: 77116
2009-07-26 11:06:11 +08:00
; int G5 = 47;
@G5 = global i32 47
; LINUX: .data
; LINUX: .globl G5
; LINUX: G5:
; LINUX: .long 47
Make the big switch: Change MCSectionMachO to represent a section *semantically* instead of syntactically as a string. This means that it keeps track of the segment, section, flags, etc directly and asmprints them in the right format. This also includes parsing and validation support for llvm-mc and "attribute(section)", so we should now start getting errors about invalid section attributes from the compiler instead of the assembler on darwin. Still todo: 1) Uniquing of darwin mcsections 2) Move all the Darwin stuff out to MCSectionMachO.[cpp|h] 3) there are a few FIXMEs, for example what is the syntax to get the S_GB_ZEROFILL segment type? llvm-svn: 78547
2009-08-10 09:39:42 +08:00
; DARWIN: .section __DATA,__data
put normal data into .data instead of .data.rel on elf systems. llvm-svn: 77116
2009-07-26 11:06:11 +08:00
; DARWIN: .globl _G5
; DARWIN: _G5:
; DARWIN: .long 47
fix PR4584 with a trivial patch now that the pieces are in place. llvm-svn: 77434
2009-07-29 13:20:33 +08:00
; PR4584
@"foo bar" = linkonce global i32 42
Move the function and data section flags into the options struct and make the functions to set them non-static. Move and rename the llvm specific backend options to avoid conflicting with the clang option. Paired with a backend commit to update. llvm-svn: 209238
2014-05-21 05:25:34 +08:00
; LINUX: .type "foo bar",@object
; LINUX: .weak "foo bar"
Remove AllowQuotesInName and friends from MCAsmInfo. Accepting quotes is a property of an assembler, not of an object file. For example, ELF can support any names for sections and symbols, but the gnu assembler only accepts quotes in some contexts and llvm-mc in a few more. LLVM should not produce different symbols based on a guess about which assembler will be reading the code it is printing. llvm-svn: 194575
2013-11-13 22:01:59 +08:00
; LINUX: "foo bar":
fix PR4584 with a trivial patch now that the pieces are in place. llvm-svn: 77434
2009-07-29 13:20:33 +08:00
Move the function and data section flags into the options struct and make the functions to set them non-static. Move and rename the llvm specific backend options to avoid conflicting with the clang option. Paired with a backend commit to update. llvm-svn: 209238
2014-05-21 05:25:34 +08:00
; DARWIN: .section __DATA,__datacoal_nt,coalesced
; DARWIN: .globl "_foo bar"
; DARWIN: .weak_definition "_foo bar"
fix PR4584 with a trivial patch now that the pieces are in place. llvm-svn: 77434
2009-07-29 13:20:33 +08:00
; DARWIN: "_foo bar":
fix PR4650: we only track sizes for certain objects, so only put something into the mergable section if it is one of our special cases. This could obviously be improved, but this is the minimal fix and restores us to the previous behavior. llvm-svn: 77679
2009-08-01 00:17:13 +08:00
; PR4650
Update tests. llvm-svn: 123591
2011-01-17 02:02:57 +08:00
@G6 = weak_odr unnamed_addr constant [1 x i8] c"\01"
fix PR4650: we only track sizes for certain objects, so only put something into the mergable section if it is one of our special cases. This could obviously be improved, but this is the minimal fix and restores us to the previous behavior. llvm-svn: 77679
2009-08-01 00:17:13 +08:00
Move the function and data section flags into the options struct and make the functions to set them non-static. Move and rename the llvm specific backend options to avoid conflicting with the clang option. Paired with a backend commit to update. llvm-svn: 209238
2014-05-21 05:25:34 +08:00
; LINUX: .type G6,@object
; LINUX: .weak G6
Various AsmWriter output cleanups. Use WriteAsOperand instead of PrintUnmangledNameSafely. llvm-svn: 78878
2009-08-13 09:36:44 +08:00
; LINUX: G6:
Move the function and data section flags into the options struct and make the functions to set them non-static. Move and rename the llvm specific backend options to avoid conflicting with the clang option. Paired with a backend commit to update. llvm-svn: 209238
2014-05-21 05:25:34 +08:00
; LINUX: .byte 1
; LINUX: .size G6, 1
fix PR4650: we only track sizes for certain objects, so only put something into the mergable section if it is one of our special cases. This could obviously be improved, but this is the minimal fix and restores us to the previous behavior. llvm-svn: 77679
2009-08-01 00:17:13 +08:00
Revert "Reapply r185872 now that the address sanitizer has been changed to support this." This reverts commit r187939. It broke an O0 build of a spec benchmark. llvm-svn: 188012
2013-08-09 05:04:16 +08:00
; DARWIN: .section __TEXT,__const_coal,coalesced
fix PR4650: we only track sizes for certain objects, so only put something into the mergable section if it is one of our special cases. This could obviously be improved, but this is the minimal fix and restores us to the previous behavior. llvm-svn: 77679
2009-08-01 00:17:13 +08:00
; DARWIN: .globl _G6
; DARWIN: .weak_definition _G6
; DARWIN:_G6:
emit .ascii and .asciz through MCStreamer. llvm-svn: 94282
2010-01-23 12:54:10 +08:00
; DARWIN: .byte 1
fix PR4650: we only track sizes for certain objects, so only put something into the mergable section if it is one of our special cases. This could obviously be improved, but this is the minimal fix and restores us to the previous behavior. llvm-svn: 77679
2009-08-01 00:17:13 +08:00
Add support emiting for 2/4 byte mergable strings to the ".rodata.str*" section on ELF targets. llvm-svn: 78066
2009-08-05 00:13:09 +08:00
Update tests. llvm-svn: 123591
2011-01-17 02:02:57 +08:00
@G7 = unnamed_addr constant [10 x i8] c"abcdefghi\00"
Add support emiting for 2/4 byte mergable strings to the ".rodata.str*" section on ELF targets. llvm-svn: 78066
2009-08-05 00:13:09 +08:00
Move the function and data section flags into the options struct and make the functions to set them non-static. Move and rename the llvm specific backend options to avoid conflicting with the clang option. Paired with a backend commit to update. llvm-svn: 209238
2014-05-21 05:25:34 +08:00
; DARWIN: __TEXT,__cstring,cstring_literals
; DARWIN: .globl _G7
Add support emiting for 2/4 byte mergable strings to the ".rodata.str*" section on ELF targets. llvm-svn: 78066
2009-08-05 00:13:09 +08:00
; DARWIN: _G7:
Move the function and data section flags into the options struct and make the functions to set them non-static. Move and rename the llvm specific backend options to avoid conflicting with the clang option. Paired with a backend commit to update. llvm-svn: 209238
2014-05-21 05:25:34 +08:00
; DARWIN: .asciz "abcdefghi"
Add support emiting for 2/4 byte mergable strings to the ".rodata.str*" section on ELF targets. llvm-svn: 78066
2009-08-05 00:13:09 +08:00
Move the function and data section flags into the options struct and make the functions to set them non-static. Move and rename the llvm specific backend options to avoid conflicting with the clang option. Paired with a backend commit to update. llvm-svn: 209238
2014-05-21 05:25:34 +08:00
; LINUX: .section .rodata.str1.1,"aMS",@progbits,1
; LINUX: .globl G7
Add support emiting for 2/4 byte mergable strings to the ".rodata.str*" section on ELF targets. llvm-svn: 78066
2009-08-05 00:13:09 +08:00
; LINUX: G7:
Move the function and data section flags into the options struct and make the functions to set them non-static. Move and rename the llvm specific backend options to avoid conflicting with the clang option. Paired with a backend commit to update. llvm-svn: 209238
2014-05-21 05:25:34 +08:00
; LINUX: .asciz "abcdefghi"
Add support emiting for 2/4 byte mergable strings to the ".rodata.str*" section on ELF targets. llvm-svn: 78066
2009-08-05 00:13:09 +08:00
Don't create multiple mergeable sections with -fdata-sections. ELF has support for sections that can be split into fixed size or null terminated entities. Since these sections can be split by the linker, it is not necessary to split them in codegen. This reduces the combined .o size in a llvm+clang build from 202,394,570 to 173,819,098 bytes. The time for linking clang with gold (on a VM, on a laptop) goes from 2.250089985 to 1.383001792 seconds. The flip side is the size of rodata in clang goes from 10,926,785 to 10,929,345 bytes. The increase seems to be because of http://sourceware.org/bugzilla/show_bug.cgi?id=17902. llvm-svn: 227431
2015-01-29 20:43:28 +08:00
; LINUX-SECTIONS: .section .rodata.str1.1,"aMS",@progbits,1
Move the function and data section flags into the options struct and make the functions to set them non-static. Move and rename the llvm specific backend options to avoid conflicting with the clang option. Paired with a backend commit to update. llvm-svn: 209238
2014-05-21 05:25:34 +08:00
; LINUX-SECTIONS: .globl G7
add llvm codegen support for -ffunction-sections and -fdata-sections, patch by Sylvere Teissier! llvm-svn: 101106
2010-04-13 08:36:43 +08:00
MC: Emit COFF section flags in the "proper" order COFF section flags are not idempotent: 'rd' will make a read-write section because 'd' implies write 'dr' will make a read-only section because 'r' disables write llvm-svn: 228490
2015-02-07 16:26:40 +08:00
; WIN32-SECTIONS: .section .rdata,"dr",one_only,_G7
Move the function and data section flags into the options struct and make the functions to set them non-static. Move and rename the llvm specific backend options to avoid conflicting with the clang option. Paired with a backend commit to update. llvm-svn: 209238
2014-05-21 05:25:34 +08:00
; WIN32-SECTIONS: .globl _G7
WinCOFF: Add support for -fdata-sections This is a pretty straight forward translation for COFF, we just need to stick the data in a COMDAT section marked as IMAGE_COMDAT_SELECT_NODUPLICATES. N.B. We must be careful to avoid sticking entities with private linkage in COMDAT groups. COFF is pretty hostile to the renaming of entities so we must be careful to disallow GlobalVariables with unstable names. llvm-svn: 204703
2014-03-25 14:14:26 +08:00
Add support emiting for 2/4 byte mergable strings to the ".rodata.str*" section on ELF targets. llvm-svn: 78066
2009-08-05 00:13:09 +08:00
Update tests. llvm-svn: 123591
2011-01-17 02:02:57 +08:00
@G8 = unnamed_addr constant [4 x i16] [ i16 1, i16 2, i16 3, i16 0 ]
Add support emiting for 2/4 byte mergable strings to the ".rodata.str*" section on ELF targets. llvm-svn: 78066
2009-08-05 00:13:09 +08:00
Move the function and data section flags into the options struct and make the functions to set them non-static. Move and rename the llvm specific backend options to avoid conflicting with the clang option. Paired with a backend commit to update. llvm-svn: 209238
2014-05-21 05:25:34 +08:00
; DARWIN: .section __TEXT,__const
; DARWIN: .globl _G8
Add support emiting for 2/4 byte mergable strings to the ".rodata.str*" section on ELF targets. llvm-svn: 78066
2009-08-05 00:13:09 +08:00
; DARWIN: _G8:
Move the function and data section flags into the options struct and make the functions to set them non-static. Move and rename the llvm specific backend options to avoid conflicting with the clang option. Paired with a backend commit to update. llvm-svn: 209238
2014-05-21 05:25:34 +08:00
; LINUX: .section .rodata.str2.2,"aMS",@progbits,2
; LINUX: .globl G8
Add support emiting for 2/4 byte mergable strings to the ".rodata.str*" section on ELF targets. llvm-svn: 78066
2009-08-05 00:13:09 +08:00
; LINUX:G8:
Update tests. llvm-svn: 123591
2011-01-17 02:02:57 +08:00
@G9 = unnamed_addr constant [4 x i32] [ i32 1, i32 2, i32 3, i32 0 ]
Add support emiting for 2/4 byte mergable strings to the ".rodata.str*" section on ELF targets. llvm-svn: 78066
2009-08-05 00:13:09 +08:00
Move the function and data section flags into the options struct and make the functions to set them non-static. Move and rename the llvm specific backend options to avoid conflicting with the clang option. Paired with a backend commit to update. llvm-svn: 209238
2014-05-21 05:25:34 +08:00
; DARWIN: .globl _G9
Add support emiting for 2/4 byte mergable strings to the ".rodata.str*" section on ELF targets. llvm-svn: 78066
2009-08-05 00:13:09 +08:00
; DARWIN: _G9:
Move the function and data section flags into the options struct and make the functions to set them non-static. Move and rename the llvm specific backend options to avoid conflicting with the clang option. Paired with a backend commit to update. llvm-svn: 209238
2014-05-21 05:25:34 +08:00
; LINUX: .section .rodata.str4.4,"aMS",@progbits,4
; LINUX: .globl G9
Add support emiting for 2/4 byte mergable strings to the ".rodata.str*" section on ELF targets. llvm-svn: 78066
2009-08-05 00:13:09 +08:00
; LINUX:G9
fix a significant difference between llvm and gcc on ELF systems: GCC would put weak zero initialized mutable data in the .bss section, we would put it into a crasy '.gnu.linkonce.b.test,"aw",@nobits' section. Fixing this will allow simplifications next up. llvm-svn: 93844
2010-01-19 11:06:01 +08:00
@G10 = weak global [100 x i32] zeroinitializer, align 32 ; <[100 x i32]*> [#uses=0]
Move the function and data section flags into the options struct and make the functions to set them non-static. Move and rename the llvm specific backend options to avoid conflicting with the clang option. Paired with a backend commit to update. llvm-svn: 209238
2014-05-21 05:25:34 +08:00
; DARWIN: .section __DATA,__datacoal_nt,coalesced
fix a significant difference between llvm and gcc on ELF systems: GCC would put weak zero initialized mutable data in the .bss section, we would put it into a crasy '.gnu.linkonce.b.test,"aw",@nobits' section. Fixing this will allow simplifications next up. llvm-svn: 93844
2010-01-19 11:06:01 +08:00
; DARWIN: .globl _G10
Move the function and data section flags into the options struct and make the functions to set them non-static. Move and rename the llvm specific backend options to avoid conflicting with the clang option. Paired with a backend commit to update. llvm-svn: 209238
2014-05-21 05:25:34 +08:00
; DARWIN: .weak_definition _G10
; DARWIN: .align 5
fix a significant difference between llvm and gcc on ELF systems: GCC would put weak zero initialized mutable data in the .bss section, we would put it into a crasy '.gnu.linkonce.b.test,"aw",@nobits' section. Fixing this will allow simplifications next up. llvm-svn: 93844
2010-01-19 11:06:01 +08:00
; DARWIN: _G10:
Move the function and data section flags into the options struct and make the functions to set them non-static. Move and rename the llvm specific backend options to avoid conflicting with the clang option. Paired with a backend commit to update. llvm-svn: 209238
2014-05-21 05:25:34 +08:00
; DARWIN: .space 400
fix a significant difference between llvm and gcc on ELF systems: GCC would put weak zero initialized mutable data in the .bss section, we would put it into a crasy '.gnu.linkonce.b.test,"aw",@nobits' section. Fixing this will allow simplifications next up. llvm-svn: 93844
2010-01-19 11:06:01 +08:00
Move the function and data section flags into the options struct and make the functions to set them non-static. Move and rename the llvm specific backend options to avoid conflicting with the clang option. Paired with a backend commit to update. llvm-svn: 209238
2014-05-21 05:25:34 +08:00
; LINUX: .bss
; LINUX: .weak G10
; LINUX: .align 32
fix a significant difference between llvm and gcc on ELF systems: GCC would put weak zero initialized mutable data in the .bss section, we would put it into a crasy '.gnu.linkonce.b.test,"aw",@nobits' section. Fixing this will allow simplifications next up. llvm-svn: 93844
2010-01-19 11:06:01 +08:00
; LINUX: G10:
Move the function and data section flags into the options struct and make the functions to set them non-static. Move and rename the llvm specific backend options to avoid conflicting with the clang option. Paired with a backend commit to update. llvm-svn: 209238
2014-05-21 05:25:34 +08:00
; LINUX: .zero 400
Add support emiting for 2/4 byte mergable strings to the ".rodata.str*" section on ELF targets. llvm-svn: 78066
2009-08-05 00:13:09 +08:00
round zero-byte .zerofill directives up to 1 byte. This should fix some "g++.dg-struct-layout-1" failures, rdar://7886017 llvm-svn: 102421
2010-04-27 15:41:44 +08:00
;; Zero sized objects should round up to 1 byte in zerofill directives.
; rdar://7886017
@G11 = global [0 x i32] zeroinitializer
@G12 = global {} zeroinitializer
@G13 = global { [0 x {}] } zeroinitializer
; DARWIN: .globl _G11
; DARWIN: .zerofill __DATA,__common,_G11,1,2
; DARWIN: .globl _G12
; DARWIN: .zerofill __DATA,__common,_G12,1,3
; DARWIN: .globl _G13
; DARWIN: .zerofill __DATA,__common,_G13,1,3
Add back r201608, r201622, r201624 and r201625 r201608 made llvm corretly handle private globals with MachO. r201622 fixed a bug in it and r201624 and r201625 were changes for using private linkage, assuming that llvm would do the right thing. They all got reverted because r201608 introduced a crash in LTO. This patch includes a fix for that. The issue was that TargetLoweringObjectFile now has to be initialized before we can mangle names of private globals. This is trivially true during the normal codegen pipeline (the asm printer does it), but LTO has to do it manually. llvm-svn: 201700
2014-02-20 01:23:20 +08:00
@G14 = private unnamed_addr constant [4 x i8] c"foo\00", align 1
; LINUX-SECTIONS: .type .LG14,@object # @G14
Don't create multiple mergeable sections with -fdata-sections. ELF has support for sections that can be split into fixed size or null terminated entities. Since these sections can be split by the linker, it is not necessary to split them in codegen. This reduces the combined .o size in a llvm+clang build from 202,394,570 to 173,819,098 bytes. The time for linking clang with gold (on a VM, on a laptop) goes from 2.250089985 to 1.383001792 seconds. The flip side is the size of rodata in clang goes from 10,926,785 to 10,929,345 bytes. The increase seems to be because of http://sourceware.org/bugzilla/show_bug.cgi?id=17902. llvm-svn: 227431
2015-01-29 20:43:28 +08:00
; LINUX-SECTIONS: .section .rodata.str1.1,"aMS",@progbits,1
Add back r201608, r201622, r201624 and r201625 r201608 made llvm corretly handle private globals with MachO. r201622 fixed a bug in it and r201624 and r201625 were changes for using private linkage, assuming that llvm would do the right thing. They all got reverted because r201608 introduced a crash in LTO. This patch includes a fix for that. The issue was that TargetLoweringObjectFile now has to be initialized before we can mangle names of private globals. This is trivially true during the normal codegen pipeline (the asm printer does it), but LTO has to do it manually. llvm-svn: 201700
2014-02-20 01:23:20 +08:00
; LINUX-SECTIONS: .LG14:
; LINUX-SECTIONS: .asciz "foo"
; LINUX-SECTIONS: .size .LG14, 4
WinCOFF: Add support for -ffunction-sections This is a pretty straight forward translation for COFF, we just need to stick the function in a COMDAT section marked as IMAGE_COMDAT_SELECT_NODUPLICATES. llvm-svn: 204565
2014-03-24 01:47:39 +08:00
COFF: Let globals with private linkage reside in their own section COFF COMDATs (for selection kinds other than 'select any') require at least one non-section symbol in the symbol table. Satisfy this by morally enhancing the linkage from private to internal. Differential Revision: http://reviews.llvm.org/D8394 llvm-svn: 232570
2015-03-18 07:54:51 +08:00
; WIN32-SECTIONS: .section .rdata,"dr",one_only,_G14
; WIN32-SECTIONS: _G14:
WinCOFF: Add support for -fdata-sections This is a pretty straight forward translation for COFF, we just need to stick the data in a COMDAT section marked as IMAGE_COMDAT_SELECT_NODUPLICATES. N.B. We must be careful to avoid sticking entities with private linkage in COMDAT groups. COFF is pretty hostile to the renaming of entities so we must be careful to disallow GlobalVariables with unstable names. llvm-svn: 204703
2014-03-25 14:14:26 +08:00
; WIN32-SECTIONS: .asciz "foo"
WinCOFF: Add support for -ffunction-sections This is a pretty straight forward translation for COFF, we just need to stick the function in a COMDAT section marked as IMAGE_COMDAT_SELECT_NODUPLICATES. llvm-svn: 204565
2014-03-24 01:47:39 +08:00
On MachO, don't put non-private constants in mergeable sections. On MachO, putting a symbol that doesn't start with a 'L' or 'l' in one of the __TEXT,__literal* sections prevents the linker from merging the context of the section. Since private GVs are the ones the get mangled to start with 'L' or 'l', we now only put those on the __TEXT,__literal* sections. llvm-svn: 216682
2014-08-29 04:13:31 +08:00
; cannot be merged on MachO, but can on other formats.
@G15 = unnamed_addr constant i64 0
; LINUX: .section .rodata.cst8,"aM",@progbits,8
; LINUX: G15:
; DARWIN: .section __TEXT,__const
; DARWIN: _G15:
; DARWIN-STATIC: .section __TEXT,__const
; DARWIN-STATIC: _G15:
; DARWIN64: .section __TEXT,__const
; DARWIN64: _G15:
Don't create multiple mergeable sections with -fdata-sections. ELF has support for sections that can be split into fixed size or null terminated entities. Since these sections can be split by the linker, it is not necessary to split them in codegen. This reduces the combined .o size in a llvm+clang build from 202,394,570 to 173,819,098 bytes. The time for linking clang with gold (on a VM, on a laptop) goes from 2.250089985 to 1.383001792 seconds. The flip side is the size of rodata in clang goes from 10,926,785 to 10,929,345 bytes. The increase seems to be because of http://sourceware.org/bugzilla/show_bug.cgi?id=17902. llvm-svn: 227431
2015-01-29 20:43:28 +08:00
; LINUX-SECTIONS: .section .rodata.cst8,"aM",@progbits,8
On MachO, don't put non-private constants in mergeable sections. On MachO, putting a symbol that doesn't start with a 'L' or 'l' in one of the __TEXT,__literal* sections prevents the linker from merging the context of the section. Since private GVs are the ones the get mangled to start with 'L' or 'l', we now only put those on the __TEXT,__literal* sections. llvm-svn: 216682
2014-08-29 04:13:31 +08:00
; LINUX-SECTIONS: G15:
MC: Emit COFF section flags in the "proper" order COFF section flags are not idempotent: 'rd' will make a read-write section because 'd' implies write 'dr' will make a read-only section because 'r' disables write llvm-svn: 228490
2015-02-07 16:26:40 +08:00
; WIN32-SECTIONS: .section .rdata,"dr",one_only,_G15
On MachO, don't put non-private constants in mergeable sections. On MachO, putting a symbol that doesn't start with a 'L' or 'l' in one of the __TEXT,__literal* sections prevents the linker from merging the context of the section. Since private GVs are the ones the get mangled to start with 'L' or 'l', we now only put those on the __TEXT,__literal* sections. llvm-svn: 216682
2014-08-29 04:13:31 +08:00
; WIN32-SECTIONS: _G15: