[commit 2/2] testsuite: code_elim.exp: Fix false FAIL

[commit 2/2] testsuite: code_elim.exp: Fix false FAIL

[Jan Kratochvil]

Hi,

with Fedora 18 x86_64 (Rawhide) there is a false FAIl since:
2167931cf70f91206efb8b01e848a55d82a347ce is the first bad commit
Author: Tom Tromey <tromey@redhat.com>
Date:   Fri May 18 15:31:39 2012 +0000
http://sourceware.org/bugzilla/show_bug.cgi?id=13907

 print &my_global_symbol
-$1 = (int *) 0x804a004
-(gdb) PASS: gdb.base/code_elim.exp: get address of my_global_symbol
+$1 = (int *) 0x804a004 <completed.5732>
+(gdb) FAIL: gdb.base/code_elim.exp: get address of my_global_symbol

Nothing really changed, just GDB now prints <completed.5732> and the testcase
expects either none or <my_global_symbol> suffix.

This is because the two symbol files (their sections .data/.bss) have
overlapped so GDB chooses randomly which symbol to display.  And since
Fedora 18 GCC puts there that "completed.5732" symbol.

This may be because add-symbol-file FILE ADDR relocates to ADDR on the .text
section and leaves other sections at their original place, which is IMO a bit
unexpected; it is a FAQ people expect more ADDR to be relocation displacement
(and not relative to .text and not limited to .text).  But this is really
outside of the scope of this testcase false FAIL.

Checked in.


Thanks,
Jan


http://sourceware.org/ml/gdb-cvs/2012-06/msg00012.html

--- src/gdb/testsuite/ChangeLog	2012/06/03 17:06:04	1.3210
+++ src/gdb/testsuite/ChangeLog	2012/06/03 17:10:28	1.3211
@@ -11,6 +11,16 @@
 	(get address of my_global_func, get address of main): Wrap them to
 	'order1' and 'order2'.
 
+	* gdb.base/code_elim.exp
+	(add-symbol-file ${testfile1} 0x100000)
+	(add-symbol-file ${testfile2} 0x200000): Wrap them to 'order1' and
+	'order2'.
+	(order1: add-symbol-file ${testfile1} 0x100000)
+	(order1: add-symbol-file ${testfile2} 0x200000)
+	(order2: add-symbol-file ${testfile2} 0x200000)
+	(order2: add-symbol-file ${testfile1} 0x100000): Add -s .data and -s
+	.bss as appropriate.
+
 2012-05-28  Jan Kratochvil  <jan.kratochvil@redhat.com>
 
 	* gdb.server/solib-list-lib.c: New file.
--- src/gdb/testsuite/gdb.base/code_elim.exp	2012/06/03 17:06:04	1.5
+++ src/gdb/testsuite/gdb.base/code_elim.exp	2012/06/03 17:10:28	1.6
@@ -128,19 +128,19 @@
 gdb_exit
 gdb_start
 
-gdb_test "add-symbol-file ${binfile1} 0x100000" \
-	"Reading symbols from .*${testfile1}\\.\\.\\.done\\." \
-	"add-symbol-file ${testfile1} 0x100000" \
-	"add symbol table from file \".*${testfile1}\" at\[ \t\r\n\]+\.text_addr = 0x100000\[\r\n\]+\\(y or n\\) " \
-	"y"
-
-gdb_test "add-symbol-file ${binfile2} 0x200000" \
-	"Reading symbols from .*${testfile2}\\.\\.\\.done\\." \
-	"add-symbol-file ${testfile2} 0x200000" \
-	"add symbol table from file \".*${testfile2}\" at\[ \t\r\n\]+\.text_addr = 0x200000\[\r\n\]+\\(y or n\\) " \
-	"y"
-
 with_test_prefix "order1" {
+    gdb_test "add-symbol-file ${binfile1} 0x100000 -s .bss 0x120000" \
+	    "Reading symbols from .*${testfile1}\\.\\.\\.done\\." \
+	    "add-symbol-file ${testfile1} 0x100000" \
+	    "add symbol table from file \".*${testfile1}\" at.*\\(y or n\\) " \
+	    "y"
+
+    gdb_test "add-symbol-file ${binfile2} 0x200000 -s .data 0x210000 -s .bss 0x220000" \
+	    "Reading symbols from .*${testfile2}\\.\\.\\.done\\." \
+	    "add-symbol-file ${testfile2} 0x200000" \
+	    "add symbol table from file \".*${testfile2}\" at.*\\(y or n\\) " \
+	    "y"
+
     not_null_var_address my_global_symbol
     not_null_var_address my_static_symbol
     not_null_var_address my_global_func
@@ -152,19 +152,19 @@
 gdb_exit
 gdb_start
 
-gdb_test "add-symbol-file ${binfile2} 0x200000" \
-	"Reading symbols from .*${testfile2}\\.\\.\\.done\\." \
-	"add-symbol-file ${testfile2} 0x200000" \
-	"add symbol table from file \".*${testfile2}\" at\[ \t\r\n\]+\.text_addr = 0x200000\[\r\n\]+\\(y or n\\) " \
-	"y"
-
-gdb_test "add-symbol-file ${binfile1} 0x100000" \
-	"Reading symbols from .*${testfile1}\\.\\.\\.done\\." \
-	"add-symbol-file ${testfile1} 0x100000" \
-	"add symbol table from file \".*${testfile1}\" at\[ \t\r\n\]+\.text_addr = 0x100000\[\r\n\]+\\(y or n\\) " \
-	"y"
-
 with_test_prefix "order2" {
+    gdb_test "add-symbol-file ${binfile2} 0x200000 -s .data 0x210000 -s .bss 0x220000" \
+	    "Reading symbols from .*${testfile2}\\.\\.\\.done\\." \
+	    "add-symbol-file ${testfile2} 0x200000" \
+	    "add symbol table from file \".*${testfile2}\" at.*\\(y or n\\) " \
+	    "y"
+
+    gdb_test "add-symbol-file ${binfile1} 0x100000 -s .bss 0x120000" \
+	    "Reading symbols from .*${testfile1}\\.\\.\\.done\\." \
+	    "add-symbol-file ${testfile1} 0x100000" \
+	    "add symbol table from file \".*${testfile1}\" at.*\\(y or n\\) " \
+	    "y"
+
     not_null_var_address my_global_symbol
     not_null_var_address my_static_symbol
     not_null_var_address my_global_func

Re: [commit 2/2] testsuite: code_elim.exp: Fix false FAIL

[Andreas Schwab]

Jan Kratochvil <jan.kratochvil@redhat.com> writes:

> +	(order2: add-symbol-file ${testfile1} 0x100000): Add -s .data and -s
> +	.bss as appropriate.

What about .sdata or .sbss (or whatever other section the compiler
chooses)?

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: [commit 2/2] testsuite: code_elim.exp: Fix false FAIL

[Jan Kratochvil]

On Sun, 03 Jun 2012 19:29:15 +0200, Andreas Schwab wrote:
> Jan Kratochvil <jan.kratochvil@redhat.com> writes:
> > +	(order2: add-symbol-file ${testfile1} 0x100000): Add -s .data and -s
> > +	.bss as appropriate.
> 
> What about .sdata or .sbss (or whatever other section the compiler
> chooses)?

As I stated I do not know how to easily do a real fix of this testcase so
I just made non-regressing improvement for the platforms I regression test.

IMO a real fix is to implement add-symbol-file to (also) accept full
displacement but this seems to be outside of the scope of this testfile/fixup
and I have too many open other work to get into another task.

If you state non-existing sections GDB will make a warning in such case.
(Sure the testcase can ignore such warning.)


Thanks,
Jan

Re: [commit 2/2] testsuite: code_elim.exp: Fix false FAIL

[Yao Qi]

On 06/04/2012 01:17 AM, Jan Kratochvil wrote:
>  with_test_prefix "order1" {
> +    gdb_test "add-symbol-file ${binfile1} 0x100000 -s .bss 0x120000" \
> +	    "Reading symbols from .*${testfile1}\\.\\.\\.done\\." \
> +	    "add-symbol-file ${testfile1} 0x100000" \
> +	    "add symbol table from file \".*${testfile1}\" at.*\\(y or n\\) " \
> +	    "y"
> +
> +    gdb_test "add-symbol-file ${binfile2} 0x200000 -s .data 0x210000 -s .bss 0x220000" \
> +	    "Reading symbols from .*${testfile2}\\.\\.\\.done\\." \
> +	    "add-symbol-file ${testfile2} 0x200000" \
> +	    "add symbol table from file \".*${testfile2}\" at.*\\(y or n\\) " \
> +	    "y"
Hi Jan,
We should first check whether '.bss' exists (by "maint info sections")
and then run this part conditionally.  We don't have '.bss' on some
ports, such as tic6x-elf.

-- 
Yao (齐尧)

Re: [commit 2/2] testsuite: code_elim.exp: Fix false FAIL

[Andreas Schwab]

Jan Kratochvil <jan.kratochvil@redhat.com> writes:

> +    gdb_test "add-symbol-file ${binfile2} 0x200000 -s .data 0x210000 -s .bss 0x220000" \
> +	    "Reading symbols from .*${testfile2}\\.\\.\\.done\\." \
> +	    "add-symbol-file ${testfile2} 0x200000" \
> +	    "add symbol table from file \".*${testfile2}\" at.*\\(y or n\\) " \
> +	    "y"
> +

add-symbol-file /home/abuild/rpmbuild/BUILD/gdb-7.4.91/gdb/testsuite/gdb.base/code_elim2 0x200000 -s .data 0x210000 -s .bss 0x220000
add symbol table from file "/home/abuild/rpmbuild/BUILD/gdb-7.4.91/gdb/testsuite/gdb.base/code_elim2" at
	.text_addr = 0x200000
	.data_addr = 0x210000
	.bss_addr = 0x220000
(y or n) y
Reading symbols from /home/abuild/rpmbuild/BUILD/gdb-7.4.91/gdb/testsuite/gdb.base/code_elim2...warning: section .data not found in /home/abuild/rpmbuild/BUILD/gdb-7.4.91/gdb/testsuite/gdb.base/code_elim2
done.
(gdb) FAIL: gdb.base/code_elim.exp: order1: add-symbol-file code_elim2 0x200000
print &my_global_symbol

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."