[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [avr-gcc-list] Link Time Optimization vs __builtin_unreachable() and
From: |
Bob Paddock |
Subject: |
Re: [avr-gcc-list] Link Time Optimization vs __builtin_unreachable() and .signature? |
Date: |
Mon, 18 Jun 2012 11:10:45 -0400 |
> Can you confirm that it is GCC that is getting rid of .signature? If so, it's
> a GCC bug and another bug report should be filed. If it is not GCC that is
> getting rid of it, but it is the linker that is doing that then it should be
> a GNU Binutils bug report. I don't know if we need to do something with the
> default linker script to tell ld to keep it, or if it is something else that
> is blowing it away.
I don't think this would be a valid bug for LTO in this case, for
either GCC or Binutils.
LTO is doing what it was meant to do as I understand it. How would it
know we are putting stuff in the .ELF file for reasons other than code
generation, unless we tell it? So the question becomes how do we tell
it?
Also noticing the -g in there, is the admonishment in the
documentation that -g and -flto are incompatible still valid?
"Link-time optimization does not work well with generation of
debugging information. Combining -flto with -g is currently
experimental and expected to produce wrong results."
I assume some debugging symbols are lost, have not looked into that yet?
I'm leaning to it is the linker, as the .lst file still has the data, see below:
85 .text
86 .section .gnu.lto_.opts,"",@progbits
87 0000 272D 6D6D
.string
"'-mmcu=atxmega128a1''-mcall-prologues''-gdwarf-2''-Os''-Werror''-pedantic''-flto''-flto-r
87 6375 3D61
87 7478 6D65
87 6761 3132
87 3861 3127
88 .text
89 .cfi_sections .debug_frame
90 .global __signature
91 .section .signature,"a",@progbits
94 __signature:
95 0000 4C .byte 76
96 0001 97 .byte -105
97 0002 1E .byte 30
98 .text
99 .Letext0:
100 .file 1
"j:\\apps\\avr-gcc-4.7.1-rc1-mingw32\\bin\\../lib/gcc/avr/4.7.1/../../../../avr/include/av
DEFINED SYMBOLS
*ABS*:00000000 signature.c
J:\Temp\ccCqYnId.s:2 *ABS*:0000003e __SP_H__
J:\Temp\ccCqYnId.s:3 *ABS*:0000003d __SP_L__
J:\Temp\ccCqYnId.s:4 *ABS*:0000003f __SREG__
J:\Temp\ccCqYnId.s:5 *ABS*:0000003b __RAMPZ__
J:\Temp\ccCqYnId.s:6 *ABS*:0000003a __RAMPY__
J:\Temp\ccCqYnId.s:7 *ABS*:00000039 __RAMPX__
J:\Temp\ccCqYnId.s:8 *ABS*:00000038 __RAMPD__
J:\Temp\ccCqYnId.s:9 *ABS*:00000034 __CCP__
J:\Temp\ccCqYnId.s:10 *ABS*:00000000 __tmp_reg__
J:\Temp\ccCqYnId.s:11 *ABS*:00000001 __zero_reg__
J:\Temp\ccCqYnId.s:94 .signature:00000000 __signature
*COM*:00000001 __gnu_lto_v1
NO UNDEFINED SYMBOLS
- [avr-gcc-list] Link Time Optimization vs __builtin_unreachable() and .signature?, Bob Paddock, 2012/06/18
- Re: [avr-gcc-list] Link Time Optimization vs __builtin_unreachable() and .signature?, Weddington, Eric, 2012/06/18
- Re: [avr-gcc-list] Link Time Optimization vs __builtin_unreachable() and .signature?,
Bob Paddock <=
- Re: [avr-gcc-list] Link Time Optimization vs __builtin_unreachable() and .signature?, Georg-Johann Lay, 2012/06/18
- Re: [avr-gcc-list] Link Time Optimization vs __builtin_unreachable() and .signature?, Bob Paddock, 2012/06/18
- Re: [avr-gcc-list] Link Time Optimization vs __builtin_unreachable() and .signature?, Georg-Johann Lay, 2012/06/18
- Re: [avr-gcc-list] Link Time Optimization vs __builtin_unreachable() and .signature?, Bob Paddock, 2012/06/19
- Re: [avr-gcc-list] Link Time Optimization vs __builtin_unreachable() and .signature?, Georg-Johann Lay, 2012/06/21
- Re: [avr-gcc-list] Link Time Optimization vs __builtin_unreachable() and .signature?, Bob Paddock, 2012/06/26
- Re: [avr-gcc-list] Link Time Optimization vs __builtin_unreachable() and .signature?, Georg-Johann Lay, 2012/06/30