This is the mail archive of the gdb-patches@sourceware.org mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: i386.record.floating.point.patch : with more testing and assurity


Could you post it in patch format?  That will make me can try it directly.  :)

Thanks,
Hui

On Mon, Jul 13, 2009 at 02:05, Michael Snyder<msnyder@vmware.com> wrote:
> paawan oza wrote:
>>
>> Hi, Thank you very much for your elaborated reply.
>> I will integrate the test example with test suite by next week-end.
>> And, also take of all yours and Hui's formatting comments.
>> I will come up with the final up to date patch with integrated test-suite.
>> Thank you again for your response. : )
>> I start working towards it.
>> Regards,
>> Oza.
>
> Excellent. ?Now let's talk about what the test should test.
>
> I don't think it is necessary to test each individual instruction
> (unless you really want to). ?We don't usually go to that extreme.
>
> What we want to establish is "does the patch do what it's intended
> to do", so let's make a statement about what it's intended to do.
> You correct me if I'm wrong, but in my impression, it can be summed
> up as: preserve and restore the state of the standard Intel floating
> point unit during record and replay".
>
> So you just need to test that the FP machine state is preserved,
> and that when we go backwards, the state is restored to what it
> was before. ?I think that means mainly the '387' registers, or
> the state that is displayed when we say "info float".
>
> To help you get started, I've attached a simple gdb test that
> checks to see if we can save and restore one register ('st0').
> This will show you in general how it is done, and if you want
> to, you can use it as a starting point, and add tests for the
> rest of the registers.
>
> The subroutines that we use, such as "gdb_test", are mostly
> defined in gdb/testsuite/lib/gdb.exp, if you need to look at
> them. ?There's some documentation in the comments, and also
> in gdb/doc/gdbint.texinfo.
>
>
>
>
>
>
>
>
> # Copyright 2009 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 3 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, see <http://www.gnu.org/licenses/>.
>
> # This file is part of the gdb testsuite.
>
> if $tracelevel {
> ? ?strace $tracelevel
> }
>
> set prms_id 0
> set bug_id 0
>
> if ![istarget "i?86-*linux*"] then {
> ? ?verbose "Skipping i387 reverse float tests."
> ? ?return
> }
>
> set testfile "i387-float-reverse"
> set srcfile ${testfile}.c
> set binfile ${objdir}/${subdir}/${testfile}
>
> # some targets have leading underscores on assembly symbols.
> # TODO: detect this automatically
> set additional_flags ""
> if [istarget "i?86-*-cygwin*"] then {
> ?set additional_flags "additional_flags=-DSYMBOL_PREFIX=\"_\""
> }
>
> if { [gdb_compile "${srcdir}/${subdir}/${srcfile}" "${binfile}" executable
> [list debug $additional_flags]] != "" } {
> ? ?untested i387-float-reverse
> ? ?return -1
> }
>
> # Get things started.
>
> gdb_exit
> gdb_start
> gdb_reinitialize_dir $srcdir/$subdir
> gdb_load ${binfile}
> runto main
>
> # Record to end of main
>
> set location [gdb_get_line_number "END I387-FLOAT-REVERSE"]
> gdb_test "record" "" "Turn on process record"
> gdb_test "until $location" ".*/$srcfile:$location.*" \
> ? ?"record to end of main"
>
> # Now rewind to beginning so we can begin testing.
>
> set location [gdb_get_line_number "BEGIN I387-FLOAT-REVERSE"]
> gdb_test "set exec-dir reverse" "" "set reverse direction"
> gdb_test "until $location" ".*/$srcfile:$location.*" \
> ? ?"rewind to beginning of main"
> gdb_test "set exec-dir forward" "" "set forward direction"
>
> # Test st0 register
>
> set location [gdb_get_line_number "test st0 register"]
> gdb_test "until $location" ".*/$srcfile:$location.*asm.*fldz.*" \
> ? ?"begin test st0"
> gdb_test "stepi" "asm.*fld1.*" ? ? ? ? ? ?"set st0 == 0"
> gdb_test "info register st0" "st0 *0\t.*" "verify st0 == 0"
> gdb_test "stepi" "asm.*nop.*" ? ? ? ? ? ? "set st0 == 1"
> gdb_test "info register st0" "st0 *1\t.*" "verify st0 == 1"
>
> # Now step backward, and check that st0 value reverts to zero.
>
> gdb_test "reverse-stepi" "asm.*fld1.*" ? ?"undo set st0 == 1"
> gdb_test "info register st0" "st0 *0\t.*" "verify st0 == 0 again"
>
>
>


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]