[patch] Fix testsuite gdb.base/prelink.exp x86 unreliability

[Jan Kratochvil <jan.kratochvil@redhat.com>]

[-- Attachment #1: Type: text/plain, Size: 274 bytes --]

Hi,

on i686 it fails in about 30% of cases when the system is already prelinked.
The problem is that prelink links the test library usually to the same place.

It is 100% PASS afterwards.  The problem is not reproducible on non-i.86 arches
(such as x86_64).


Regards,
Jan

[-- Attachment #2: gdb-prelink-i386.patch --]
[-- Type: text/plain, Size: 1450 bytes --]

2008-03-21  Jan Kratochvil  <jan.kratochvil@redhat.com>

	Fix random false FAILs on i386.
	* gdb.base/prelink.exp: Use `--no-exec-shield' for prelink.

--- ./gdb/testsuite/gdb.base/prelink.exp	1 Jan 2008 22:53:19 -0000	1.7
+++ ./gdb/testsuite/gdb.base/prelink.exp	21 Mar 2008 20:09:59 -0000
@@ -47,7 +47,15 @@ if { [gdb_compile "${srcdir}/${subdir}/$
     return -1
 }
 
-if {[catch "system \"prelink -qNR ${libfile}\""] != 0} {
+# `--no-exec-shield' is for i386 where prelink in the exec-shield mode is
+# forced to push all the libraries tight together to fit into the first two
+# memory areas (either the ASCII Shield area or at least below the executable).
+# In this case its -R option cannot be applied and we falsely FAIL here as if
+# the system is already prelinked prelink has no choice how to randomize the
+# single new unprelinked library address without wasting the first one/two
+# memory areas.  We do not care of the efficiency of loading such resulting
+# exec-shield unfriendly prelinked library.
+if {[catch "system \"prelink -qNR --no-exec-shield ${libfile}\""] != 0} {
     # Maybe we don't have prelink.
     return -1
 }
@@ -92,7 +100,7 @@ if {[catch "system \"prelink -uN ${libfi
     untested "${testfile}.so was not prelinked, maybe system libraries are not prelinked?"
     return 0
 }
-catch "system \"prelink -qNR ${libfile}\""
+catch "system \"prelink -qNR --no-exec-shield ${libfile}\""
 
 # Start with a fresh gdb