[Bug libgcc/56300] New: Add __sync_fetch_and_add_8 and other 8-byte atomic functions to 32-bit MIPS targets
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56300 Bug #: 56300 Summary: Add __sync_fetch_and_add_8 and other 8-byte atomic functions to 32-bit MIPS targets Classification: Unclassified Product: gcc Version: unknown Status: UNCONFIRMED Severity: normal Priority: P3 Component: libgcc AssignedTo: unassig...@gcc.gnu.org ReportedBy: chaoyin...@gcc.gnu.org There is no hardware support for 8 bytes atomic operations on 32-bit MIPS targets. Unfortunately, programmers still want to use them. Ex: please check this webkit bug: https://bugs.webkit.org/show_bug.cgi?id=106739 and the following example. Ex: fu@debian6:~/dev/test$ cat sync.c long long K; void main() { __sync_fetch_and_add(K, 1); } fu@debian6:~/dev/test$ gcc -o sync sync.c /tmp/ccUoSg9M.o: In function `main': sync.c:(.text+0x28): undefined reference to `__sync_fetch_and_add_8' collect2: ld returned 1 exit status That will be great, if we can add missing atomic functions (via pthread implementation or other methods) to libgcc for 32-bit MIPS targets. Thanks a lot!
[Bug libgcc/56300] Add __sync_fetch_and_add_8 and other 8-byte atomic functions to 32-bit MIPS targets
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56300 --- Comment #2 from chaoyingfu at gcc dot gnu.org chaoyingfu at gcc dot gnu.org 2013-02-12 22:21:09 UTC --- (In reply to comment #1) Use libatomic in 4.8 to fix this. Good to know this. Thanks!
[Bug rtl-optimization/54369] Delayed-branch pass in reorg.c removes too many instructions
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54369 --- Comment #3 from chaoyingfu at gcc dot gnu.org chaoyingfu at gcc dot gnu.org 2012-08-31 00:38:39 UTC --- The fall-through path of this jump has a barrier (resulted from __builtin_unreachable), and the target of this jump has a return instruction. The algorithm in reorg.c may be more conservative to keep this jump for code correctness. Just another solution to this issue. Thanks!
[Bug rtl-optimization/54369] New: Delayed-branch pass in reorg.c removes too many instructions
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54369 Bug #: 54369 Summary: Delayed-branch pass in reorg.c removes too many instructions Classification: Unclassified Product: gcc Version: unknown Status: UNCONFIRMED Severity: normal Priority: P3 Component: rtl-optimization AssignedTo: unassig...@gcc.gnu.org ReportedBy: chaoyin...@gcc.gnu.org Created attachment 28079 -- http://gcc.gnu.org/bugzilla/attachment.cgi?id=28079 A simple test for an issue in reorg.c Please check this example when using GCC 4.6 for mips-linux-gnu. (I think the issue is in GCC 4.7 and later as well. But due to the basic block ordering, I cannot reproduce this issue.) Ex: # /gcc46branch/build-mips/gcc/cc1plus -quiet Bug.i -o Bug.s -mips32r2 -O2 # cat Bug.s _Z3bugP9ValueRTRK: .frame $sp,0,$31 # vars= 0, regs= 0/0, args= 0, gp= 0 .mask 0x,0 .fmask 0x,0 .setnoreorder .setnomacro lw $2,0($4) - lw $2,0($2) - (NO RETURN IN THIS FUNCTION.) .setmacro .setreorder .end_Z3bugP9ValueRTRK .cfi_endproc $LFE13: .size _Z3bugP9ValueRTRK, .-_Z3bugP9ValueRTRK .ident GCC: (GNU) 4.6.4 20120823 (prerelease) [gcc-4_6-branch revision 190613] The delayed-branch pass in reorg.c is too aggressive, such that a return instruction and others are removed. From debugging code in reorg.c: delete_jump() - delete_computation() - delete_related_insn(), I think we should not use delete_related_insn() that removes branches and all the following instructions. A simple fix may be just using delete_insn() in delete_computation(). Ex: # svn diff reorg.c Index: reorg.c === --- reorg.c (revision 190636) +++ reorg.c (working copy) @@ -3298,7 +3298,7 @@ delete_prior_computation (note, insn); } - delete_related_insns (insn); + delete_insn (insn); } /* If all INSN does is set the pc, delete it, Ex: (Bug.s after this fix.) # cat Bug.s _Z3bugP9ValueRTRK: .frame $sp,0,$31 # vars= 0, regs= 0/0, args= 0, gp= 0 .mask 0x,0 .fmask 0x,0 .setnoreorder .setnomacro lw $2,0($4) lw $2,0($2) $L4: j $31 - OK srl $2,$2,8 - OK .setmacro .setreorder .end_Z3bugP9ValueRTRK .cfi_endproc $LFE13: .size _Z3bugP9ValueRTRK, .-_Z3bugP9ValueRTRK .ident GCC: (GNU) 4.6.4 20120823 (prerelease) [gcc-4_6-branch revision 190613] Thanks!
[Bug bootstrap/54209] [4.8 Regression] Failed to build gcc for Android/x86
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54209 --- Comment #2 from chaoyingfu at gcc dot gnu.org chaoyingfu at gcc dot gnu.org 2012-08-09 17:23:40 UTC --- MIPS provides a version of link.h in Android NDK as follows: Ex: From android-ndk-r8b/platforms/android-9/arch-mips/usr/include# cat link.h /* For building unwind-dw2-fde-glibc.c for MIPS frame unwinding, we need to have link.h that defines struct dl_phdr_info, ELFW(type), and dl_iterate_phdr(). */ #include sys/types.h #include elf.h struct dl_phdr_info { Elf32_Addr dlpi_addr; const char *dlpi_name; const Elf32_Phdr *dlpi_phdr; Elf32_Half dlpi_phnum; }; #if _MIPS_SZPTR == 32 #define ElfW(type) Elf32_##type #elif _MIPS_SZPTR == 64 #define ElfW(type) Elf64_##type #endif int dl_iterate_phdr(int (*cb)(struct dl_phdr_info *info, size_t size, void *data), void *data); For x86, you can create link.h as well. Or we can guard this define with MIPS targets. Ex: #if !defined(inhibit_libc) defined(HAVE_LD_EH_FRAME_HDR) \ defined(__BIONIC__) defined(__mips__) # define USE_PT_GNU_EH_FRAME #endif Thanks!
[Bug bootstrap/54209] [4.8 Regression] Failed to build gcc for Android/x86
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54209 --- Comment #4 from chaoyingfu at gcc dot gnu.org chaoyingfu at gcc dot gnu.org 2012-08-09 18:45:41 UTC --- (In reply to comment #3) (In reply to comment #2) MIPS provides a version of link.h in Android NDK as follows: Ex: From android-ndk-r8b/platforms/android-9/arch-mips/usr/include# cat link.h /* For building unwind-dw2-fde-glibc.c for MIPS frame unwinding, we need to have link.h that defines struct dl_phdr_info, ELFW(type), and dl_iterate_phdr(). */ #include sys/types.h #include elf.h struct dl_phdr_info { Elf32_Addr dlpi_addr; const char *dlpi_name; const Elf32_Phdr *dlpi_phdr; Elf32_Half dlpi_phnum; }; #if _MIPS_SZPTR == 32 #define ElfW(type) Elf32_##type #elif _MIPS_SZPTR == 64 #define ElfW(type) Elf64_##type #endif int dl_iterate_phdr(int (*cb)(struct dl_phdr_info *info, size_t size, void *data), void *data); For x86, you can create link.h as well. Or we can guard this define with MIPS targets. Why isn't link.h in AOSP Bionic C library? ARM doesn't use eh_frame, so there is no need to create link.h at the beginning for the Android project, I guess. For MIPS, we create our own link.h to work with eh_frame unwinding. Thanks!