[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Bug gas/19520] [2.26 regression] R_386_GOT32X relocation breaks gcc boo
From: |
ro at CeBiTec dot Uni-Bielefeld.DE |
Subject: |
[Bug gas/19520] [2.26 regression] R_386_GOT32X relocation breaks gcc bootstrap with non-gld/gold linker |
Date: |
Wed, 27 Jan 2016 14:00:29 +0000 |
https://sourceware.org/bugzilla/show_bug.cgi?id=19520
--- Comment #7 from Rainer Orth <ro at CeBiTec dot Uni-Bielefeld.DE> ---
> --- Comment #4 from H.J. Lu <hjl.tools at gmail dot com> ---
> Please checkout users/hjl/pr19520/master branch. I added a
> configure-time option, --enable-new-x86-relocations, and a
> run-time option, -mnew-relocations=.
I've successfully bootstrapped gcc mainline with gas 2.26 with your
patch applied and /bin/ld.
A couple of comments, though:
* As Andreas noted, today's new is tomorrow's old. The option needs a
more descriptive name.
* It needs documentation in gas/doc/c-i386.texi, perhaps together with
the new configure option.
* I wonder if the gas option should be called -mnew-relocations=(yes|no)
or -m(no-)new-relocations. There seems to be precedent for both
versions.
* There's a superfluous trailing ] in gas/configure, both for the new
option and --enable-compressed-debug-sections (copy-and-paste error,
no doubt):
@@ -1415,6 +1416,8 @@ Optional Features:
--enable-checking enable run-time checks
--enable-compressed-debug-sections={all,gas,none}
compress debug sections by default]
+ --enable-new-x86-relocations
+ generate new x86 relocations by default]
--enable-werror treat compile warnings as errors
--enable-build-warnings enable build-time compiler warnings
--disable-nls do not use Native Language Support
* Typo in gas/configure.ac:
diff --git a/gas/configure.ac b/gas/configure.ac
--- a/gas/configure.ac
+++ b/gas/configure.ac
@@ -77,6 +77,20 @@ AC_ARG_ENABLE(compressed_debug_sections,
[...]
+AC_DEFINE_UNQUOTED(DEFAULT_GENERATE_X86_NEW_RELOCATIONS,
+ $ac_default_new_x86_relocations,
+ [Define to 1 if you want generate new x86 relocations by default.])
^ to generate
* I believe --enable-new-x86-relocations should default to 0 on
i?86-*-solaris2.[0-9], i?86-*-solaris2.1[01], x86_64-*-solaris2.1[01]
and to 1 on i?86-solaris2.1[2-9], x86_64-*-solaris2.1[2-9] so this
works out of the box.
Rainer
--
You are receiving this mail because:
You are on the CC list for the bug.
- [Bug gas/19520] New: [2.26 regression] R_386_GOT32X relocation breaks gcc bootstrap with non-gld/gold linker, ro at TechFak dot Uni-Bielefeld.DE, 2016/01/26
- [Bug gas/19520] [2.26 regression] R_386_GOT32X relocation breaks gcc bootstrap with non-gld/gold linker, hjl.tools at gmail dot com, 2016/01/26
- [Bug gas/19520] [2.26 regression] R_386_GOT32X relocation breaks gcc bootstrap with non-gld/gold linker, ro at CeBiTec dot Uni-Bielefeld.DE, 2016/01/26
- [Bug gas/19520] [2.26 regression] R_386_GOT32X relocation breaks gcc bootstrap with non-gld/gold linker, hjl.tools at gmail dot com, 2016/01/26
- [Bug gas/19520] [2.26 regression] R_386_GOT32X relocation breaks gcc bootstrap with non-gld/gold linker, hjl.tools at gmail dot com, 2016/01/26
- [Bug gas/19520] [2.26 regression] R_386_GOT32X relocation breaks gcc bootstrap with non-gld/gold linker, address@hidden, 2016/01/27
- [Bug gas/19520] [2.26 regression] R_386_GOT32X relocation breaks gcc bootstrap with non-gld/gold linker, hjl.tools at gmail dot com, 2016/01/27
- [Bug gas/19520] [2.26 regression] R_386_GOT32X relocation breaks gcc bootstrap with non-gld/gold linker,
ro at CeBiTec dot Uni-Bielefeld.DE <=
- [Bug gas/19520] [2.26 regression] R_386_GOT32X relocation breaks gcc bootstrap with non-gld/gold linker, hjl.tools at gmail dot com, 2016/01/27
- [Bug gas/19520] [2.26 regression] R_386_GOT32X relocation breaks gcc bootstrap with non-gld/gold linker, richard at netbsd dot org, 2016/01/31
- [Bug gas/19520] [2.26 regression] R_386_GOT32X relocation breaks gcc bootstrap with non-gld/gold linker, hjl.tools at gmail dot com, 2016/01/31