Re: [patch] New testcase for PR:1291

Re: [patch] New testcase for PR:1291

[Michael Elizabeth Chastain]

> A) move it and make it generic.
> B) leave it and write it in assembly.

I like (B) better.  The backtrace code is all about specific
assembly instructions.  So I'd like it better in the style of the i386
prologue tests, with explicit assembly.

Michael C

Re: [patch] New testcase for PR:1291

[Kris Warkentin]

> > A) move it and make it generic.
> > B) leave it and write it in assembly.
>
> I like (B) better.  The backtrace code is all about specific
> assembly instructions.  So I'd like it better in the style of the i386
> prologue tests, with explicit assembly.
>
> Michael C

Okay so this bit of assembly adequately demonstrates the problem.  There is
a bit of an issue here though that perhaps some sh4 expert can clear up for
me.  If I run this and go "until sub2", it stops on sub2 and works fine.
If, however, I go "until sub1", followed by "until sub2", it stops a couple
instructions before sub2 and I have to 'si' into it.  I can modify the test
case to not bother checking sub1 (since we know it's okay anyway) but I'm
still puzzled as to why this is having the problem.

cheers,

Kris


void sub1 (void);
void sub2 (void);

main()
{
        sub1();
        sub2();
}

asm(".text\n"
    "    .align 5\n"
    "sub1:\n"
    "    mov.l  r14,@-r15\n"
    "    add    #-128,r15\n"
    "    add    #-128,r15\n"
    "    mov    r15,r14\n"
    "    mov.w  .STACK1,r7\n"
    "    add    r7,r14\n"
    "    mov    r14,r15\n"
    "    mov.l  @r15+,r14\n"
    "    rts\n"
    "    nop\n"
    "    .align 1\n"
    ".STACK1:\n"
    "    .short  256\n");

asm(".text\n"
    "    .align 5\n"
    "sub2:\n"
    "    mov.l  r14,@-r15\n"
    "    mov.w  .STACK2,r3\n"
    "    sub    r3,r15\n"
    "    mov    r15,r14\n"
    "    mov.w  .STACK2,r7\n"
    "    add    r7,r14\n"
    "    mov    r14,r15\n"
    "    mov.l  @r15+,r14\n"
    "    rts\n"
    "    nop\n"
    "    .align 1\n"
    ".STACK2:\n"
    "    .short  260\n");

Re: [patch] New testcase for PR:1291

[Elena Zannoni]

Kris Warkentin writes:
 > > > A) move it and make it generic.
 > > > B) leave it and write it in assembly.
 > >
 > > I like (B) better.  The backtrace code is all about specific
 > > assembly instructions.  So I'd like it better in the style of the i386
 > > prologue tests, with explicit assembly.
 > >
 > > Michael C
 > 
 > Okay so this bit of assembly adequately demonstrates the problem.  There is
 > a bit of an issue here though that perhaps some sh4 expert can clear up for
 > me.  If I run this and go "until sub2", it stops on sub2 and works fine.
 > If, however, I go "until sub1", followed by "until sub2", it stops a couple
 > instructions before sub2 and I have to 'si' into it.  I can modify the test
 > case to not bother checking sub1 (since we know it's okay anyway) but I'm
 > still puzzled as to why this is having the problem.
 > 
 > cheers,
 > 
 > Kris
 > 

Ok.  Not sure why that happens. It defeinitely seems like a bug. File
a pr, and KFAIL it in the testsuite, with the pr number.

elena


 > 
 > void sub1 (void);
 > void sub2 (void);
 > 
 > main()
 > {
 >         sub1();
 >         sub2();
 > }
 > 
 > asm(".text\n"
 >     "    .align 5\n"
 >     "sub1:\n"
 >     "    mov.l  r14,@-r15\n"
 >     "    add    #-128,r15\n"
 >     "    add    #-128,r15\n"
 >     "    mov    r15,r14\n"
 >     "    mov.w  .STACK1,r7\n"
 >     "    add    r7,r14\n"
 >     "    mov    r14,r15\n"
 >     "    mov.l  @r15+,r14\n"
 >     "    rts\n"
 >     "    nop\n"
 >     "    .align 1\n"
 >     ".STACK1:\n"
 >     "    .short  256\n");
 > 
 > asm(".text\n"
 >     "    .align 5\n"
 >     "sub2:\n"
 >     "    mov.l  r14,@-r15\n"
 >     "    mov.w  .STACK2,r3\n"
 >     "    sub    r3,r15\n"
 >     "    mov    r15,r14\n"
 >     "    mov.w  .STACK2,r7\n"
 >     "    add    r7,r14\n"
 >     "    mov    r14,r15\n"
 >     "    mov.l  @r15+,r14\n"
 >     "    rts\n"
 >     "    nop\n"
 >     "    .align 1\n"
 >     ".STACK2:\n"
 >     "    .short  260\n");
 > 
 > 

Re: [patch] New testcase for PR:1291

[Kris Warkentin]

You'll have to explain that to me a bit better.  KFAIL?  Does that mean I
leave the test as is and mark it as an expected failure somehow?  Or do I
modify the test and make a note?

cheers,

Kris

----- Original Message ----- 
From: "Elena Zannoni" <ezannoni@redhat.com>
To: "Kris Warkentin" <kewarken@qnx.com>
Cc: "Michael Elizabeth Chastain" <mec@shout.net>; <drow@mvista.com>;
<gdb-patches@sources.redhat.com>
Sent: Tuesday, October 28, 2003 11:02 AM
Subject: Re: [patch] New testcase for PR:1291


> Kris Warkentin writes:
>  > > > A) move it and make it generic.
>  > > > B) leave it and write it in assembly.
>  > >
>  > > I like (B) better.  The backtrace code is all about specific
>  > > assembly instructions.  So I'd like it better in the style of the
i386
>  > > prologue tests, with explicit assembly.
>  > >
>  > > Michael C
>  >
>  > Okay so this bit of assembly adequately demonstrates the problem.
There is
>  > a bit of an issue here though that perhaps some sh4 expert can clear up
for
>  > me.  If I run this and go "until sub2", it stops on sub2 and works
fine.
>  > If, however, I go "until sub1", followed by "until sub2", it stops a
couple
>  > instructions before sub2 and I have to 'si' into it.  I can modify the
test
>  > case to not bother checking sub1 (since we know it's okay anyway) but
I'm
>  > still puzzled as to why this is having the problem.
>  >
>  > cheers,
>  >
>  > Kris
>  >
>
> Ok.  Not sure why that happens. It defeinitely seems like a bug. File
> a pr, and KFAIL it in the testsuite, with the pr number.
>
> elena
>
>
>  >
>  > void sub1 (void);
>  > void sub2 (void);
>  >
>  > main()
>  > {
>  >         sub1();
>  >         sub2();
>  > }
>  >
>  > asm(".text\n"
>  >     "    .align 5\n"
>  >     "sub1:\n"
>  >     "    mov.l  r14,@-r15\n"
>  >     "    add    #-128,r15\n"
>  >     "    add    #-128,r15\n"
>  >     "    mov    r15,r14\n"
>  >     "    mov.w  .STACK1,r7\n"
>  >     "    add    r7,r14\n"
>  >     "    mov    r14,r15\n"
>  >     "    mov.l  @r15+,r14\n"
>  >     "    rts\n"
>  >     "    nop\n"
>  >     "    .align 1\n"
>  >     ".STACK1:\n"
>  >     "    .short  256\n");
>  >
>  > asm(".text\n"
>  >     "    .align 5\n"
>  >     "sub2:\n"
>  >     "    mov.l  r14,@-r15\n"
>  >     "    mov.w  .STACK2,r3\n"
>  >     "    sub    r3,r15\n"
>  >     "    mov    r15,r14\n"
>  >     "    mov.w  .STACK2,r7\n"
>  >     "    add    r7,r14\n"
>  >     "    mov    r14,r15\n"
>  >     "    mov.l  @r15+,r14\n"
>  >     "    rts\n"
>  >     "    nop\n"
>  >     "    .align 1\n"
>  >     ".STACK2:\n"
>  >     "    .short  260\n");
>  >
>  >
>

Re: [patch] New testcase for PR:1291

[Daniel Jacobowitz]

On Tue, Oct 28, 2003 at 11:37:50AM -0500, Kris Warkentin wrote:
> You'll have to explain that to me a bit better.  KFAIL?  Does that mean I
> leave the test as is and mark it as an expected failure somehow?  Or do I
> modify the test and make a note?

It's just like xfail.  Grep for kfail in the testsuite for examples;
kfail takes as argument a PR number (file it in GNATS).

> 
> cheers,
> 
> Kris
> 
> ----- Original Message ----- 
> From: "Elena Zannoni" <ezannoni@redhat.com>
> To: "Kris Warkentin" <kewarken@qnx.com>
> Cc: "Michael Elizabeth Chastain" <mec@shout.net>; <drow@mvista.com>;
> <gdb-patches@sources.redhat.com>
> Sent: Tuesday, October 28, 2003 11:02 AM
> Subject: Re: [patch] New testcase for PR:1291
> 
> 
> > Kris Warkentin writes:
> >  > > > A) move it and make it generic.
> >  > > > B) leave it and write it in assembly.
> >  > >
> >  > > I like (B) better.  The backtrace code is all about specific
> >  > > assembly instructions.  So I'd like it better in the style of the
> i386
> >  > > prologue tests, with explicit assembly.
> >  > >
> >  > > Michael C
> >  >
> >  > Okay so this bit of assembly adequately demonstrates the problem.
> There is
> >  > a bit of an issue here though that perhaps some sh4 expert can clear up
> for
> >  > me.  If I run this and go "until sub2", it stops on sub2 and works
> fine.
> >  > If, however, I go "until sub1", followed by "until sub2", it stops a
> couple
> >  > instructions before sub2 and I have to 'si' into it.  I can modify the
> test
> >  > case to not bother checking sub1 (since we know it's okay anyway) but
> I'm
> >  > still puzzled as to why this is having the problem.
> >  >
> >  > cheers,
> >  >
> >  > Kris
> >  >
> >
> > Ok.  Not sure why that happens. It defeinitely seems like a bug. File
> > a pr, and KFAIL it in the testsuite, with the pr number.
> >
> > elena
> >
> >
> >  >
> >  > void sub1 (void);
> >  > void sub2 (void);
> >  >
> >  > main()
> >  > {
> >  >         sub1();
> >  >         sub2();
> >  > }
> >  >
> >  > asm(".text\n"
> >  >     "    .align 5\n"
> >  >     "sub1:\n"
> >  >     "    mov.l  r14,@-r15\n"
> >  >     "    add    #-128,r15\n"
> >  >     "    add    #-128,r15\n"
> >  >     "    mov    r15,r14\n"
> >  >     "    mov.w  .STACK1,r7\n"
> >  >     "    add    r7,r14\n"
> >  >     "    mov    r14,r15\n"
> >  >     "    mov.l  @r15+,r14\n"
> >  >     "    rts\n"
> >  >     "    nop\n"
> >  >     "    .align 1\n"
> >  >     ".STACK1:\n"
> >  >     "    .short  256\n");
> >  >
> >  > asm(".text\n"
> >  >     "    .align 5\n"
> >  >     "sub2:\n"
> >  >     "    mov.l  r14,@-r15\n"
> >  >     "    mov.w  .STACK2,r3\n"
> >  >     "    sub    r3,r15\n"
> >  >     "    mov    r15,r14\n"
> >  >     "    mov.w  .STACK2,r7\n"
> >  >     "    add    r7,r14\n"
> >  >     "    mov    r14,r15\n"
> >  >     "    mov.l  @r15+,r14\n"
> >  >     "    rts\n"
> >  >     "    nop\n"
> >  >     "    .align 1\n"
> >  >     ".STACK2:\n"
> >  >     "    .short  260\n");
> >  >
> >  >
> >
> 
> 
> 

-- 
Daniel Jacobowitz
MontaVista Software                         Debian GNU/Linux Developer

Re: [patch] New testcase for PR:1291

[Michael Elizabeth Chastain]

Copyright nagging: gdb1291.c needs either a copyright notice,
or an explicit notice that copyright has been abandoned and
the file is in the public domain.

(gcc has a policy that they want bug reports to include
pre-processed C source, which means all the comments were gone.
But we have no such policy.  Instead, we have a policy that all
files longer than 10 lines must have copyright notices,
including test files).

Also, Andrew says that the preferred incantation is
"this file is part of gdb", not "this file is part of the gdb
test suite".

Other than that, I don't have any stylistic complaints.
Lightly proofread.  Not tested.  Busy day for me.  :(

Michael C

Changelog:

2003-10-24  Kris Warkentin  <kewarken@qnx.com>

 * gdb.arch/gdb1291.c: New test file.
 * gdb.arch/gdb1291.exp: New test script.

[patch] New testcase for PR:1291

[Kris Warkentin]

Changelog:

2003-10-24  Kris Warkentin  <kewarken@qnx.com>

 * gdb.arch/gdb1291.c: New test file.
 * gdb.arch/gdb1291.exp: New test script.

-------gdb1291.c--------
sub1()
{
        int buf[64];
}

sub2()
{
        int buf[65];
}

main()
{
        sub1();
        sub2();
}

-----gdb1291.exp-----

# Copyright 2003 Free Software Foundation, Inc.

# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 2 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software
# Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.

# Please email any bugs, comments, and/or additions to this file to:
# bug-gdb@gnu.org

# This file is part of the gdb testsuite.

# Tests for PR:1291.  Ensure that backtrace works properly for stack
# frames greater than 256 bytes.

if $tracelevel {
    strace $tracelevel
}

# Test SH backtraces with >256 byte frame stack. (PR:1291)

set prms_id 0
set bug_id 0

if ![istarget "sh-*-*"] then {
    verbose "Skipping SH backtrace tests."
    return
}

set testfile "gdb1291"
set srcfile ${testfile}.c
set binfile ${objdir}/${subdir}/${testfile}
if { [gdb_compile "${srcdir}/${subdir}/${srcfile}" "${binfile}" executable
{debug}] != "" } {
    gdb_suppress_entire_file "Testcase compile failed, so all tests in this
file will automatically fail."
}

gdb_exit
gdb_start
gdb_reinitialize_dir $srcdir/$subdir
gdb_load ${binfile}

#
# Run to `main' where we begin our tests.
#

if ![runto_main] then {
    gdb_suppress_tests
}

gdb_test "u sub1" "sub1*" "get to sub1"
gdb_test "bt" "#0\[ \t\]*sub1 \\(\\) at sh-bt.*\r\n#1\[ \t\]*$hex in main.*"
\
 "backtrace in gdb1291"

gdb_test "u sub2" "sub2*" "get to sub2"
gdb_test "bt" "#0\[ \t\]*sub2 \\(\\) at sh-bt.*\r\n#1\[ \t\]*$hex in main.*"
\
 "backtrace in gdb1291"

Re: [patch] New testcase for PR:1291

[Daniel Jacobowitz]

On Fri, Oct 24, 2003 at 11:41:50AM -0400, Kris Warkentin wrote:
> Changelog:
> 
> 2003-10-24  Kris Warkentin  <kewarken@qnx.com>
> 
>  * gdb.arch/gdb1291.c: New test file.
>  * gdb.arch/gdb1291.exp: New test script.
> 
> -------gdb1291.c--------
> sub1()
> {
>         int buf[64];
> }
> 
> sub2()
> {
>         int buf[65];
> }
> 
> main()
> {
>         sub1();
>         sub2();
> }

A test which looks like this should not, I think, go in gdb.arch or be
skipped for non-SH.  What you may want to do is leave it where it is
and write the testcase in assembly - that's what the gdb.arch directory
is for.  See i386-prolouge.c.

> 
> -----gdb1291.exp-----
> 
> # Copyright 2003 Free Software Foundation, Inc.
> 
> # This program is free software; you can redistribute it and/or modify
> # it under the terms of the GNU General Public License as published by
> # the Free Software Foundation; either version 2 of the License, or
> # (at your option) any later version.
> #
> # This program is distributed in the hope that it will be useful,
> # but WITHOUT ANY WARRANTY; without even the implied warranty of
> # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> # GNU General Public License for more details.
> #
> # You should have received a copy of the GNU General Public License
> # along with this program; if not, write to the Free Software
> # Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
> 
> # Please email any bugs, comments, and/or additions to this file to:
> # bug-gdb@gnu.org
> 
> # This file is part of the gdb testsuite.
> 
> # Tests for PR:1291.  Ensure that backtrace works properly for stack
> # frames greater than 256 bytes.
> 
> if $tracelevel {
>     strace $tracelevel
> }
> 
> # Test SH backtraces with >256 byte frame stack. (PR:1291)
> 
> set prms_id 0
> set bug_id 0
> 
> if ![istarget "sh-*-*"] then {
>     verbose "Skipping SH backtrace tests."
>     return
> }
> 
> set testfile "gdb1291"
> set srcfile ${testfile}.c
> set binfile ${objdir}/${subdir}/${testfile}
> if { [gdb_compile "${srcdir}/${subdir}/${srcfile}" "${binfile}" executable
> {debug}] != "" } {
>     gdb_suppress_entire_file "Testcase compile failed, so all tests in this
> file will automatically fail."
> }
> 
> gdb_exit
> gdb_start
> gdb_reinitialize_dir $srcdir/$subdir
> gdb_load ${binfile}
> 
> #
> # Run to `main' where we begin our tests.
> #
> 
> if ![runto_main] then {
>     gdb_suppress_tests
> }
> 
> gdb_test "u sub1" "sub1*" "get to sub1"
> gdb_test "bt" "#0\[ \t\]*sub1 \\(\\) at sh-bt.*\r\n#1\[ \t\]*$hex in main.*"
> \
>  "backtrace in gdb1291"
> 
> gdb_test "u sub2" "sub2*" "get to sub2"
> gdb_test "bt" "#0\[ \t\]*sub2 \\(\\) at sh-bt.*\r\n#1\[ \t\]*$hex in main.*"
> \
>  "backtrace in gdb1291"
> 
> 
> 

-- 
Daniel Jacobowitz
MontaVista Software                         Debian GNU/Linux Developer

Re: [patch] New testcase for PR:1291

[Kris Warkentin]

> > main()
> > {
> >         sub1();
> >         sub2();
> > }
> 
> A test which looks like this should not, I think, go in gdb.arch or be
> skipped for non-SH.  What you may want to do is leave it where it is
> and write the testcase in assembly - that's what the gdb.arch directory
> is for.  See i386-prolouge.c.

I think I need an executive decision here.

A) move it and make it generic.
B) leave it and write it in assembly.

Pick one.

cheers,

Kris

Re: [patch] New testcase for PR:1291

[Daniel Jacobowitz]

On Fri, Oct 24, 2003 at 01:37:52PM -0400, Kris Warkentin wrote:
> > > main()
> > > {
> > >         sub1();
> > >         sub2();
> > > }
> > 
> > A test which looks like this should not, I think, go in gdb.arch or be
> > skipped for non-SH.  What you may want to do is leave it where it is
> > and write the testcase in assembly - that's what the gdb.arch directory
> > is for.  See i386-prolouge.c.
> 
> I think I need an executive decision here.
> 
> A) move it and make it generic.
> B) leave it and write it in assembly.
> 
> Pick one.

Either's fine with me; Elena liked the test you posted so go ahead. 
Just stick it in gdb.base?

-- 
Daniel Jacobowitz
MontaVista Software                         Debian GNU/Linux Developer

Re: [patch] New testcase for PR:1291

[Elena Zannoni]

Kris Warkentin writes:
 > Changelog:
 > 
 > 2003-10-24  Kris Warkentin  <kewarken@qnx.com>
 > 
 >  * gdb.arch/gdb1291.c: New test file.
 >  * gdb.arch/gdb1291.exp: New test script.

Thanks!
MichaelC, OK with you?

elena


 > 
 > -------gdb1291.c--------
 > sub1()
 > {
 >         int buf[64];
 > }
 > 
 > sub2()
 > {
 >         int buf[65];
 > }
 > 
 > main()
 > {
 >         sub1();
 >         sub2();
 > }
 > 
 > -----gdb1291.exp-----
 > 
 > # Copyright 2003 Free Software Foundation, Inc.
 > 
 > # This program is free software; you can redistribute it and/or modify
 > # it under the terms of the GNU General Public License as published by
 > # the Free Software Foundation; either version 2 of the License, or
 > # (at your option) any later version.
 > #
 > # This program is distributed in the hope that it will be useful,
 > # but WITHOUT ANY WARRANTY; without even the implied warranty of
 > # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 > # GNU General Public License for more details.
 > #
 > # You should have received a copy of the GNU General Public License
 > # along with this program; if not, write to the Free Software
 > # Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
 > 
 > # Please email any bugs, comments, and/or additions to this file to:
 > # bug-gdb@gnu.org
 > 
 > # This file is part of the gdb testsuite.
 > 
 > # Tests for PR:1291.  Ensure that backtrace works properly for stack
 > # frames greater than 256 bytes.
 > 
 > if $tracelevel {
 >     strace $tracelevel
 > }
 > 
 > # Test SH backtraces with >256 byte frame stack. (PR:1291)
 > 
 > set prms_id 0
 > set bug_id 0
 > 
 > if ![istarget "sh-*-*"] then {
 >     verbose "Skipping SH backtrace tests."
 >     return
 > }
 > 
 > set testfile "gdb1291"
 > set srcfile ${testfile}.c
 > set binfile ${objdir}/${subdir}/${testfile}
 > if { [gdb_compile "${srcdir}/${subdir}/${srcfile}" "${binfile}" executable
 > {debug}] != "" } {
 >     gdb_suppress_entire_file "Testcase compile failed, so all tests in this
 > file will automatically fail."
 > }
 > 
 > gdb_exit
 > gdb_start
 > gdb_reinitialize_dir $srcdir/$subdir
 > gdb_load ${binfile}
 > 
 > #
 > # Run to `main' where we begin our tests.
 > #
 > 
 > if ![runto_main] then {
 >     gdb_suppress_tests
 > }
 > 
 > gdb_test "u sub1" "sub1*" "get to sub1"
 > gdb_test "bt" "#0\[ \t\]*sub1 \\(\\) at sh-bt.*\r\n#1\[ \t\]*$hex in main.*"
 > \
 >  "backtrace in gdb1291"
 > 
 > gdb_test "u sub2" "sub2*" "get to sub2"
 > gdb_test "bt" "#0\[ \t\]*sub2 \\(\\) at sh-bt.*\r\n#1\[ \t\]*$hex in main.*"
 > \
 >  "backtrace in gdb1291"
 >