From: Alan Modra <amodra@gmail.com>
To: binutils@sourceware.org
Subject: Fix testsuite XPASSes
Date: Fri, 13 May 2011 04:14:00 -0000 [thread overview]
Message-ID: <20110513041419.GB30815@bubble.grove.modra.org> (raw)
This cures some XPASSes. In the case of the gas redef2 and redef3
tests, certain targets will now FAIL that silently XFAILed before. If
my analysis is correct these new FAILs are real failures: performing a
final link on the object file resulted in a binary with wrong
addresses.
binutils/testsuite/
* binutils-all/objcopy.exp objcopy_text): Remove xfails for sh-rtems
and tic4x.
gas/testsuite/
* gas/all/gas.exp: Remove some xfails on redef2 and redef3 tests.
Update comments.
* gas/hppa/unsorted/unsorted.exp: Run globalbug test on appropriate
targets rather than xfailing.
ld/testsuite/
* ld-elf/flags1.d: Don't xfail tic6x.
* ld-scripts/weak.exp: Correct comments. Don't xfail sh-pe. Remove
redundant xfail.
* ld-selective/sel-dump.exp: Don't xfail alpha.
* ld-selective/selective.exp: Run for alpha.
* ld-undefined/weak-undef.exp: Don't xfail hppa64.
Index: binutils/testsuite/binutils-all/objcopy.exp
===================================================================
RCS file: /cvs/src/src/binutils/testsuite/binutils-all/objcopy.exp,v
retrieving revision 1.72
diff -u -p -r1.72 objcopy.exp
--- binutils/testsuite/binutils-all/objcopy.exp 6 Apr 2011 17:09:50 -0000 1.72
+++ binutils/testsuite/binutils-all/objcopy.exp 13 May 2011 02:30:43 -0000
@@ -87,8 +87,8 @@ proc objcopy_test {testname srcfile} {
setup_xfail "m68*-*-sysv*" "m68*-apple-aux*"
setup_xfail "m8*-*"
setup_xfail "or32-*-rtems*" "or32-*-coff"
- setup_xfail "sh-*-coff*" "sh-*-rtems*"
- setup_xfail "tic4x-*-*" "tic80-*-*" "w65-*"
+ setup_xfail "sh-*-coff*"
+ setup_xfail "tic80-*-*" "w65-*"
clear_xfail "hppa*64*-*-hpux*" "hppa*-*-linux*" "hppa*-*-lites*"
clear_xfail "hppa*-*-*n*bsd*" "hppa*-*-rtems*" "*-*-*elf*"
Index: gas/testsuite/gas/all/gas.exp
===================================================================
RCS file: /cvs/src/src/gas/testsuite/gas/all/gas.exp,v
retrieving revision 1.72
diff -u -p -r1.72 gas.exp
--- gas/testsuite/gas/all/gas.exp 18 Apr 2011 13:43:20 -0000 1.72
+++ gas/testsuite/gas/all/gas.exp 13 May 2011 02:30:49 -0000
@@ -130,20 +130,22 @@ case $target_triplet in {
{ z80-*-* } { }
default {
run_dump_test redef
- # The next two tests can fail if the target does not convert fixups
- # against ordinary symbols into relocations against section symbols.
- # This is usually revealed by the error message:
- # symbol `sym' required but not present
- setup_xfail "*arm*-*-*aout*" "*arm*-*-*coff" \
- "*arm*-*-pe" "m68hc*-*-*" \
- "rx-*-*" "vax*-*-*" "z8k-*-*"
+ # These targets fail redef2 because they disallow redefined
+ # symbols on relocs.
+ setup_xfail "m68hc*-*-*" "rx-*-*" "vax*-*-*" "z8k-*-*"
run_dump_test redef2
- setup_xfail "*-*-aix*" "bfin-*-*" "hppa*-*-hpux*" \
- "i\[3-7\]86-*-*coff" "i\[3-7\]86-*-*pe" "i\[3-7\]86-*-go32*" \
- "i\[3-7\]86-*-cygwin*" "i\[3-7\]86-*-mingw*" \
- "x86_64-*-*coff" "x86_64-*-*pe" \
- "x86_64-*-cygwin*" "x86_64-*-mingw*" \
- "m68hc*-*-*" "or32-*-*" "rx-*-*" "vax*-*-*" "z8k-*-*"
+ setup_xfail "m68hc*-*-*" "rx-*-*" "vax*-*-*" "z8k-*-*"
+ # rs6000-aix disallows redefinition via .comm.
+ setup_xfail "*-*-aix*"
+ # SOM uses a different syntax for .comm
+ setup_xfail "hppa*-*-hpux*"
+ # These targets fail redef3 because section contents for the
+ # word referencing the .comm sym is not zero and/or its reloc
+ # has a non-zero addend. Relaxing the test would hide real
+ # failures such as or32-elf.
+ setup_xfail "bfin-*-*" "i\[3-7\]86-*-*coff" \
+ "i\[3-7\]86-*-*pe" "i\[3-7\]86-*-go32*" \
+ "i\[3-7\]86-*-cygwin*" "i\[3-7\]86-*-mingw*" "86_64-*-mingw*"
run_dump_test redef3
gas_test_error "redef4.s" "" ".set for symbol already used as label"
gas_test_error "redef5.s" "" ".set for symbol already defined through .comm"
Index: gas/testsuite/gas/hppa/unsorted/unsorted.exp
===================================================================
RCS file: /cvs/src/src/gas/testsuite/gas/hppa/unsorted/unsorted.exp,v
retrieving revision 1.12
diff -u -p -r1.12 unsorted.exp
--- gas/testsuite/gas/hppa/unsorted/unsorted.exp 2 Sep 2009 07:24:24 -0000 1.12
+++ gas/testsuite/gas/hppa/unsorted/unsorted.exp 13 May 2011 02:30:50 -0000
@@ -245,11 +245,13 @@ if [istarget hppa*-*-*] then {
# Test bug where switching between subspaces creates bogus alignments
do_align4_test
- # Test a problem where $global$ is defined, then used within the
- # same source file.
- setup_xfail hppa*-*-*
- gas_test "globalbug.s" "" "" "Use \$global\$ in file which defines it"
-
+ if { ([istarget *-*-osf*] || [istarget *-*-hpux*] || [istarget *-*-bsd*]
+ || [istarget *-*-mpeix*] || [istarget *-*-hiux*])
+ && ![istarget hppa*64*-*-hpux11*] } {
+ # Test a problem where $global$ is defined, then used within the
+ # same source file.
+ gas_test "globalbug.s" "" "" "Use \$global\$ in file which defines it"
+ }
# Test that importing a defined symbol doesn't screw up the symbol's
# space/subspace.
do_import_test
Index: ld/testsuite/ld-elf/flags1.d
===================================================================
RCS file: /cvs/src/src/ld/testsuite/ld-elf/flags1.d,v
retrieving revision 1.7
diff -u -p -r1.7 flags1.d
--- ld/testsuite/ld-elf/flags1.d 28 Oct 2010 06:21:49 -0000 1.7
+++ ld/testsuite/ld-elf/flags1.d 13 May 2011 02:31:07 -0000
@@ -3,10 +3,10 @@
#objcopy_linked_file: --set-section-flags .post_text_reserve=contents,alloc,load,readonly,code
#readelf: -l --wide
#xfail: "avr-*-*" "dlx-*-*" "h8300-*-*" "i960-*-*" "ip2k-*-*" "m32r-*-*"
-#xfail: "moxie-*-*" "mt-*-*" "msp430-*-*" "tic6x-*-*"
+#xfail: "moxie-*-*" "mt-*-*" "msp430-*-*"
#xfail: "*-*-hpux*" "hppa*64*-*-*"
-# Fails on the AVR, DLX, H8300, I960, IP2K, M32R, MOXIE, MT, MSP430 and
-# TI C6X because the two sections are not merged into one segment.
+# Fails on the AVR, DLX, H8300, I960, IP2K, M32R, MOXIE, MT, and MSP430
+# because the two sections are not merged into one segment.
# (There is no good reason why they have to be).
# Fails on HPUX systems because the .type pseudo-op behaves differently.
# Fails on hppa64 because a PHDR is always added.
Index: ld/testsuite/ld-scripts/weak.exp
===================================================================
RCS file: /cvs/src/src/ld/testsuite/ld-scripts/weak.exp,v
retrieving revision 1.16
diff -u -p -r1.16 weak.exp
--- ld/testsuite/ld-scripts/weak.exp 2 Sep 2009 07:25:41 -0000 1.16
+++ ld/testsuite/ld-scripts/weak.exp 13 May 2011 02:31:08 -0000
@@ -22,17 +22,16 @@
set testname "weak symbols"
-# This test only works for ELF targets.
+# This test only works for ELF targets and some PE targets.
# It ought to work for some a.out targets, but it doesn't.
if {! [is_elf_format] && ! [is_pecoff_format]} {
unsupported $testname
return
}
-# Weak symbols are broken for non-i386 PE targets.
-if {! [istarget i?86-*-*]} {
+# Weak symbols are broken for most PE targets.
+if {! [istarget i?86-*-*] && ! [istarget sh-*-*]} {
setup_xfail *-*-pe*
- setup_xfail x86_64-*-pe*
}
# hppa64 and or32 are incredibly broken
Index: ld/testsuite/ld-selective/sel-dump.exp
===================================================================
RCS file: /cvs/src/src/ld/testsuite/ld-selective/sel-dump.exp,v
retrieving revision 1.11
diff -u -p -r1.11 sel-dump.exp
--- ld/testsuite/ld-selective/sel-dump.exp 2 Sep 2009 07:25:41 -0000 1.11
+++ ld/testsuite/ld-selective/sel-dump.exp 13 May 2011 02:31:08 -0000
@@ -28,7 +28,7 @@ set test_list [lsort [glob -nocomplain $
for { set i 0 } { $i < [llength $test_list] } { incr i } {
# We need to strip the ".d", but can leave the dirname.
verbose [file rootname [lindex $test_list $i]]
- setup_xfail "alpha*-*" "am33*-*" "arc*-*" "d30v*-*" "dlx*-*"
+ setup_xfail "am33*-*" "arc*-*" "d30v*-*" "dlx*-*"
setup_xfail "hppa*64-*-*" "i370*-*" "i860*-*" "i960*-*" "ia64*-*"
setup_xfail "m88*-*" "mn10200-*" "mep-*" "or32-*" "pj-*"
run_dump_test [file rootname [lindex $test_list $i]]
Index: ld/testsuite/ld-selective/selective.exp
===================================================================
RCS file: /cvs/src/src/ld/testsuite/ld-selective/selective.exp,v
retrieving revision 1.44
diff -u -p -r1.44 selective.exp
--- ld/testsuite/ld-selective/selective.exp 6 Apr 2011 17:09:56 -0000 1.44
+++ ld/testsuite/ld-selective/selective.exp 13 May 2011 02:31:08 -0000
@@ -28,7 +28,7 @@ if ![is_elf_format] {
}
# These targets do not support selective linking
-if {[istarget "alpha*-*-*"] || [istarget "am33*-*-*"] ||
+if {[istarget "am33*-*-*"] ||
[istarget "arc-*-*"] || [istarget "d30v-*-*"] ||
[istarget "dlx-*-*"] || [istarget "hppa*64*-*-*"] ||
[istarget "i370-*-*"] || [istarget "i860-*-*"] ||
Index: ld/testsuite/ld-undefined/weak-undef.exp
===================================================================
RCS file: /cvs/src/src/ld/testsuite/ld-undefined/weak-undef.exp,v
retrieving revision 1.11
diff -u -p -r1.11 weak-undef.exp
--- ld/testsuite/ld-undefined/weak-undef.exp 8 Nov 2010 06:43:40 -0000 1.11
+++ ld/testsuite/ld-undefined/weak-undef.exp 13 May 2011 02:31:08 -0000
@@ -50,7 +50,7 @@ if {! [istarget i?86-*-*]} {
setup_xfail *-*-pe*
}
-setup_xfail hppa64-*-* pj-*-*
+setup_xfail pj-*-*
if {! [ld_assemble $as $srcdir/$subdir/weak-undef.s tmpdir/weak-undef.o]} then {
# It's OK if .weak doesn't work on this target.
--
Alan Modra
Australia Development Lab, IBM
next reply other threads:[~2011-05-13 4:14 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-05-13 4:14 Alan Modra [this message]
2011-05-14 0:52 ` Alan Modra
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20110513041419.GB30815@bubble.grove.modra.org \
--to=amodra@gmail.com \
--cc=binutils@sourceware.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).