packages: binutils/binutils.spec, binutils/binutils-fix-i686-kernel.patch (...

baggins baggins at pld-linux.org
Wed May 2 20:18:22 CEST 2012


Author: baggins                      Date: Wed May  2 18:18:22 2012 GMT
Module: packages                      Tag: HEAD
---- Log message:
- looks like gcc 4.7 "fix" breaks kernel builds on i686

---- Files affected:
packages/binutils:
   binutils.spec (1.359 -> 1.360) , binutils-fix-i686-kernel.patch (NONE -> 1.1)  (NEW)

---- Diffs:

================================================================
Index: packages/binutils/binutils.spec
diff -u packages/binutils/binutils.spec:1.359 packages/binutils/binutils.spec:1.360
--- packages/binutils/binutils.spec:1.359	Thu Apr 26 21:34:32 2012
+++ packages/binutils/binutils.spec	Wed May  2 20:18:17 2012
@@ -38,6 +38,7 @@
 Patch7:		%{name}-libtool-m.patch
 Patch8:		%{name}-build-id.patch
 Patch9:		%{name}-tooldir.patch
+Patch10:	%{name}-fix-i686-kernel.patch
 URL:		http://sources.redhat.com/binutils/
 BuildRequires:	autoconf >= 2.64
 BuildRequires:	automake >= 1:1.11
@@ -160,6 +161,7 @@
 %patch7 -p1
 %patch8 -p0
 %patch9 -p1
+%patch10 -p1
 
 # file contains hacks for ac 2.59 only
 %{__rm} config/override.m4
@@ -360,6 +362,9 @@
 All persons listed below can be reached at <cvs_login>@pld-linux.org
 
 $Log$
+Revision 1.360  2012/05/02 18:18:17  baggins
+- looks like gcc 4.7 "fix" breaks kernel builds on i686
+
 Revision 1.359  2012/04/26 19:34:32  qboosh
 - tests require statis -lstdc++ now
 

================================================================
Index: packages/binutils/binutils-fix-i686-kernel.patch
diff -u /dev/null packages/binutils/binutils-fix-i686-kernel.patch:1.1
--- /dev/null	Wed May  2 20:18:22 2012
+++ packages/binutils/binutils-fix-i686-kernel.patch	Wed May  2 20:18:17 2012
@@ -0,0 +1,32 @@
+It seems that this:
+http://sourceware.org/bugzilla/show_bug.cgi?id=13621
+breaks i686 3.3.x kernel build
+
+--- binutils-2.22.52.0.2/bfd/linker.c~	2012-04-25 19:42:51.000000000 +0200
++++ binutils-2.22.52.0.2/bfd/linker.c	2012-05-02 18:37:51.247665432 +0200
+@@ -3198,10 +3198,12 @@
+ 	best = prev;
+     }
+ 
+-  /* Refuse to choose a section for which we are out of bounds.  */
+-  /* ??? This may make most of the above moot.  */
+-  if (addr < best->vma || addr > best->vma + best->size)
+-    best = bfd_abs_section_ptr;
++// This code causes kernel issues in i686
++// http://lists.pld-linux.org/mailman/pipermail/pld-devel-pl/2012-April/155404.html
++//  /* Refuse to choose a section for which we are out of bounds.  */
++//  /* ??? This may make most of the above moot.  */
++//  if (addr < best->vma || addr > best->vma + best->size)
++//    best = bfd_abs_section_ptr;
+ 
+   return best;
+ }
+--- binutils-2.22.52.0.2/ld/testsuite/ld-elf/warn2.d~	2012-04-25 19:42:51.000000000 +0200
++++ binutils-2.22.52.0.2/ld/testsuite/ld-elf/warn2.d	2012-05-02 19:43:00.852579191 +0200
+@@ -13,5 +13,5 @@
+ # construct and that the symbol still appears as expected.
+ 
+ #...
+- +[0-9]+: +[0-9a-f]+ +20 +OBJECT +GLOBAL +DEFAULT +ABS Foo
++ +[0-9]+: +[0-9a-f]+ +20 +OBJECT +GLOBAL +DEFAULT +[1-9] Foo
+ #pass
================================================================

---- CVS-web:
    http://cvs.pld-linux.org/packages/binutils/binutils.spec?r1=1.359&r2=1.360



More information about the pld-cvs-commit mailing list